Nothing Here Is Live. Here Is Exactly What Would Change That.

Ignore-Until-Trigger Edition — Saturday, August 1, 2026

Three companies on this list are finished as equity stories. Eight more sit under an acquirer or a recent IPO, where the math gets capped before anybody drafts a job description. The remaining eighty are running on a clock.

Brex stopped being a pre-IPO equity story on April 7, the day Capital One closed. Wiz stopped on March 11, when Google closed. Cruise stopped in February 2025, when GM took full ownership and pointed the program away from robotaxis. Nothing reopens the original thesis at any of the three. What can open is a different job entirely: a mandate inside the acquirer that names the acquired surface. Different equity math, and typically six to eighteen months out.

Everything else is a queue waiting to mature, which is the whole premise of this list.

If you set only five alerts this month

Corrected scores, ranked by expected value.

1. Ramp — a new requisition ID on the Ashby board. Company 10, role 13. No other entry with a confirmed, funded design seat behind it scores higher.
2. Snowflake — successor search. The only clean design-leadership departure on the recognized board, and departure-to-successor is the fastest-moving mechanism there is.
3. Anthropic — any design-adjacent requisition on the careers board. Company 11, top of the list. The tell is a design engineer or design ops posting. It is never the leadership role.
4. One geography title search. "Staff Product Designer" plus "San Francisco" or "Remote — US," scoped to Databricks, Datadog, Headway, and Suno. Four funded mandates, one alert. When the location field changes there is zero lead time.
5. Viz.ai — role 14, the highest role score anywhere here, one company-rubric point off Act. Board unverified. Verify once, then set the alert and forget it.

The best five minutes you will spend this week is Mercury's careers page. Company 10, craft culture, unverified board.

Your monitor budget is twenty minutes a week

An ignore list you read costs you a Saturday. An ignore list you monitor costs you ten alert conditions. Here is how I would spend the twenty, ranked by yield per minute.

One exception inside that two-minute block. Cluster 7 holds three of the eight highest role scores on the list, so a spare minute goes there.

What moved

• Routed out of this list. These carry live, directly verifiable senior design mandates and belong in your active queue: Stripe, Abridge, Gusto, Amplitude, HackerOne, Atlassian, Samsara, Linear, Harvey, Instacart, Faire, Chime, Robinhood, Okta, Vanta, Maven Clinic, Ambience, Zip, Babylist, Capital One. Several were sitting on the last ignore list. That is the system doing its job.
• Downgraded on decay. Ramp (Director, Product Design, 54 days at cutoff). Salesforce (VP Product Design, Informatica, 78 days). Rubrik (February record, no confirmed repost). Adobe (prior Director pages marked filled). Brex went down on capital structure rather than posting age.
• Hot elsewhere, not here. DocuSign, CrowdStrike, and Webflow each named a new product executive inside the last thirty days. Active queue.

A correction I owe you

I have been giving you a thirty-to-ninety-day rule for the gap between a new chief product officer landing and a design leadership requisition appearing. I built it on one company. Amplitude's CPO arrived April 14, Wave launched June 10, the Head of Product Design posted July 7.

One sequence is not a study. I went looking for a validated basis and could not find one — not for CPO-to-posting, not for departure-to-successor, not for junior-cluster-to-leadership-search. Treat the rule as pattern-inferred. Good enough to set a monitor by. Not good enough to reorganize your week around.

The validated material is narrower and more useful anyway. A peer-reviewed field experiment in Management Science found that employers clear the first batch of applications quickly, that applicants did better when fewer competitors were already queued, and that among observed interview cases the first interview came a median of six days after first application. Ashby's platform data puts application volume roughly twice as high in a posting's first week. Neither isolates senior design leadership. I would rather flag that than let you build on it.

Reading the evidence labels

Absence is the load-bearing claim in a list like this. A false "nothing here" walks you past an open door, so I label the epistemics instead of smoothing them over.

One note on tiers, because it looks wrong otherwise. The archetype tracks the evaluation pattern, not the sector. Healthcare/Regulated gets applied to defense, autonomous vehicles, and regulated fintech anywhere the panel will grade you on high-consequence decision design, compliance constraint, and failure modes. Where you see (high-consequence evaluation), that's why.

Role scores are scored against the qualifying requisition as it would look if the trigger fires. Read them as forecasts. Every entry here is Ignore-until-trigger, and each trigger names a specific board, a specific requisition type, or a specific person.

1. The capital event already closed it

Cleanest mechanism on the board, and the one cluster where patience is the wrong reflex. Acquisition does not decay an equity thesis. It ends it. High confidence on every event below; each links to a filing, an official completion record, or, for Armis, independent trade reporting of the close. What survives is a mandate inside the acquirer naming the acquired surface, and that usually waits on integration to produce a product boundary first.

1. Brex. Corporate cards, expense management, and treasury for startups and mid-market finance teams.

• Stage / Size / Capital: Acquired; formerly late-stage private. A Capital One property now.
• Offices: San Francisco; parent HQ McLean, VA.
• Value chain: Downstream financial application. Value now sits in Capital One's commercial banking stack, which means a bank's governance model sets design scope.
• Best at: Enforcing spend policy at the moment of purchase instead of at month-end reconciliation.
• Signals: No recent signal — Capital One completed the acquisition April 7, 2026. One reconciliation you need: Capital One sits in your active queue on a live Director, Product Design plus Accelerator and Concierge mandates. I read all three. None name Brex surfaces. The trigger has not fired.
• Tier: Enterprise platform.
• Rubric: Company 6/12 — AI 2, Stage/equity 1, Design influence 2, Trajectory 1. Role 9/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 1.
• Portfolio match: The live surface inside a bank is the spend exception. An employee's transaction breaks a rule and somebody approves, denies, or amends it inside a few seconds. Structurally that is Carrier IQ's quote-review gate at Agentic Labs, where automated output had to be stoppable without breaking the automation. Differentiator: most fintech design candidates have never had to make an override read as an ordinary step in a workflow rather than an error state.
• Verdict: Ignore-until-trigger. Trigger: a Capital One requisition naming Brex product surfaces, stating the reporting line, and disclosing long-term incentive structure. No equity disclosure, no opening, whatever the title says.

2. Wiz. Agentless cloud security posture management. Scans cloud environments and ranks exploitable attack paths.

• Stage / Size / Capital: Acquired; part of Google Cloud.
• Offices: New York, Tel Aviv; Google Cloud presence in Sunnyvale.
• Value chain: Midstream security platform folded into a hyperscaler. Value concentrates upstream at Google Cloud, which caps independent product authority.
• Best at: Collapsing thousands of cloud misconfigurations into a ranked list of paths an attacker could actually walk.
• Signals: No recent signal — Google completed the acquisition March 11, 2026; board state not certified this pass.
• Tier: Enterprise platform.
• Rubric: Company 6/12 — AI 2, Stage/equity 1, Design influence 2, Trajectory 1. Role 8/15 — Comp 2, Scope 1, Craft 2, AI exposure 2, Portfolio 1.
• Portfolio match: The attack-path graph is a ranking-explanation problem. Why is this path first, and what happens to the order if I dismiss it. Your Alibaba search relevance sprint is the nearest published analogue: reordering results people already distrusted, and making the reorder legible. Differentiator: you have shipped a ranked-list redesign against a measured behavior change (+20% daily transactions), not a visual refresh.
• Verdict: Ignore-until-trigger. Trigger: a Google Cloud Staff+ product design requisition explicitly owning Wiz surfaces, not "cloud security" in the abstract.

3. Moveworks. Enterprise AI assistant that resolves IT and HR support requests through chat.

• Stage / Size / Capital: Acquired; part of ServiceNow.
• Offices: Mountain View; parent HQ Santa Clara.
• Value chain: Midstream workflow automation sitting on enterprise systems of record. Post-close, value concentrates in ServiceNow's platform.
• Best at: Closing a support ticket without a human ever seeing it, and knowing when not to try.
• Signals: No recent signal — ServiceNow completed the acquisition December 15, 2025.
• Tier: Enterprise platform.
• Rubric: Company 6/12 — AI 3, Stage/equity 1, Design influence 1, Trajectory 1. Role 8/15 — Comp 2, Scope 1, Craft 2, AI exposure 2, Portfolio 1.
• Portfolio match: Unsolved surface: agent-to-human handoff. What the employee sees when the assistant quits mid-resolution. That is the second handoff in "Trust Is the New Interface," by name. Differentiator: you have a published, citable model for this specific failure, which is rarer than agent experience in general.
• Verdict: Ignore-until-trigger. Trigger: a ServiceNow requisition naming Moveworks or EmployeeWorks conversation, search, or escalation ownership at Staff+ or Director level.

4. Confluent. Managed Apache Kafka. Real-time data streaming infrastructure for enterprises.

• Stage / Size / Capital: Formerly public; acquired by IBM.
• Offices: Mountain View; parent HQ Armonk, NY.
• Value chain: Upstream data infrastructure. Design influence at this layer is structurally low. The buyer is a platform engineer and the interface is mostly configuration.
• Best at: Making streaming data operable for teams who have no interest in running Kafka clusters.
• Signals: No recent signal — IBM disclosed completion March 17, 2026; board state not certified this pass.
• Tier: Enterprise platform.
• Rubric: Company 5/12 — AI 1, Stage/equity 1, Design influence 1, Trajectory 2. Role 7/15 — Comp 2, Scope 1, Craft 2, AI exposure 1, Portfolio 1.
• Portfolio match: Weak. Saying so plainly. The only bridge is pipeline observability for agent systems, which you handle daily at TinyFish through agent traces, and you have no published design proof at this layer. This buyer wants proof.
• Verdict: Ignore-until-trigger. Skip it unless the trigger fires with explicit agent-governance scope. Trigger: an IBM requisition naming Confluent streaming and agent governance in one mandate, Bay Area or US-remote.

5. Informatica. Enterprise data management, cataloging, and master data governance.

• Stage / Size / Capital: Acquired; part of Salesforce.
• Offices: Redwood City.
• Value chain: Midstream data governance now feeding Salesforce's agent platform. Value concentrates where governed data meets agent action.
• Best at: Knowing where enterprise data came from and who is permitted to use it.
• Signals: No recent signal — Salesforce completed the acquisition November 18, 2025; the standing VP Product Design page for Informatica is stale (see Cluster 3, entry 9).
• Tier: Enterprise platform.
• Rubric: Company 7/12 — AI 2, Stage/equity 1, Design influence 2, Trajectory 2. Role 9/15 — Comp 3, Scope 2, Craft 1, AI exposure 2, Portfolio 1.
• Portfolio match: Post-integration, the job is making provenance visible at the instant an agent touches a record. Which data, whose permission, what it means if the lineage is stale. Thermo Fisher mySupply is your closest proof: six pharma partners on one governed supply surface, 100% adoption, because every party could see the chain of custody. Differentiator: you have shipped governance-visible workflows where the audit trail was the product rather than a compliance afterthought.
• Verdict: Ignore-until-trigger. Trigger: a new Salesforce requisition number tying Informatica lineage to Agentforce, replacing the stale May page.

6. Armis. Asset intelligence and exposure management for unmanaged, operational, and medical devices.

• Stage / Size / Capital: Acquired; part of ServiceNow.
• Offices: San Francisco, Tel Aviv.
• Value chain: Midstream security telemetry feeding an enterprise workflow platform. Design scope depends entirely on whether ServiceNow gives the exposure surface its own console.
• Best at: Seeing devices nobody installed an agent on. Infusion pumps, industrial controllers, badge readers.
• Signals: No recent signal — ServiceNow completed the acquisition April 20, 2026.
• Tier: Healthcare/Regulated (high-consequence evaluation).
• Rubric: Company 7/12 — AI 2, Stage/equity 1, Design influence 2, Trajectory 2. Role 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 2.
• Portfolio match: The hard case is a device you cannot patch and cannot take offline, flagged vulnerable at 2 a.m. What does the operator do next? Retail Velocity's field-alert triage solved the same structure: high-volume machine alerts routed to people who could act on a handful of them. Differentiator: you have designed alert triage where the binding constraint was human capacity, not detection accuracy.
• Verdict: Ignore-until-trigger. Trigger: a Staff+ or Director design requisition at ServiceNow naming Armis operational-technology or medical-device exposure.

7. Veza. Identity and access governance. Maps who and what can reach which data across cloud systems.

• Stage / Size / Capital: Acquired; part of ServiceNow.
• Offices: Redwood City.
• Value chain: Midstream identity control plane. One of the few security layers where AI agents are now first-class subjects, which lifts the ceiling if the acquirer treats it as a product instead of a feature.
• Best at: Answering "can this identity reach this data" across systems never built to answer it together.
• Signals: No recent signal — ServiceNow completed the acquisition March 2, 2026.
• Tier: Enterprise platform.
• Rubric: Company 7/12 — AI 2, Stage/equity 1, Design influence 2, Trajectory 2. Role 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 3, Portfolio 1.
• Portfolio match: Granting an autonomous agent standing access to production data is the sharpest version of the delegation problem there is. Your trust ladder (Watch, then Verify, then Delegate) maps straight onto scoping agent permissions over time rather than at provisioning. Differentiator: you run that calibration in production at TinyFish today, so the framework holds up when someone pushes on it.
• Verdict: Ignore-until-trigger. Trigger: a Staff+ or Director role naming identity controls for AI agents specifically, not human identity governance broadly.

8. Navan. Corporate travel booking and expense management on one platform.

• Stage / Size / Capital: Public; IPO completed October 31, 2025.
• Offices: Palo Alto, San Francisco.
• Value chain: Downstream application on travel inventory and card rails. Value concentrates at the booking-and-policy decision, which design owns.
• Best at: Making policy compliance the easiest path during booking instead of a rejection after the fact.
• Signals: No recent signal — IPO completed October 31, 2025; board state not certified this pass.
• Tier: Growth-stage platform.
• Rubric: Company 7/12 — AI 2, Stage/equity 2, Design influence 2, Trajectory 1. Role 9/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 1.
• Portfolio match: Navan's agentic ambition breaks the first time an assistant books something the traveler never intended, which makes reversibility the design problem. Allē transfers: two parties, one transaction, a redemption flow correctable without a support call, 3.2x redemption across 30M members. Differentiator: dual-surface reversibility with a commercial number attached is uncommon in a portfolio.
• Verdict: Ignore-until-trigger. The equity window narrowed at IPO. Trigger: a Staff+ requisition owning agentic booking with disclosed refresh equity.

9. Klarna. Buy-now-pay-later payments plus a shopping app with AI-driven merchant discovery.

• Stage / Size / Capital: Public; IPO completed September 10, 2025 at $40 per share.
• Offices: Stockholm HQ; US presence including New York and Columbus.
• Value chain: Downstream consumer finance with a merchant network. Value concentrates at the credit decision, which is heavily regulated and heavily designed.
• Best at: Underwriting a consumer at checkout in the time a page takes to load.
• Signals: No recent signal — IPO completed September 10, 2025; US design hiring geography not certified this pass.
• Tier: Healthcare/Regulated (high-consequence evaluation).
• Rubric: Company 7/12 — AI 2, Stage/equity 2, Design influence 2, Trajectory 1. Role 9/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 1.
• Portfolio match: Klarna's exposed nerve is adverse-action explanation. Telling a non-expert why an automated system declined them, in language a regulator accepts and a person does not experience as punishment. American Red Cross is the proof, and the transferable decision is the wording of a negative outcome. $847K disbursed to people in crisis, where the constraint was making a refusal comprehensible instead of humiliating. Differentiator: you have designed the decline case, not just the happy path, for users with money on the line.
• Verdict: Ignore-until-trigger. Trigger: a US-eligible Staff+ or Director product design requisition with AI decision-surface scope.

10. Groq. Custom inference chips and a cloud service for fast large-language-model serving.

• Stage / Size / Capital: Private; $650M in growth capital announced June 22, 2026.
• Offices: Mountain View.
• Value chain: Upstream inference infrastructure. Design influence here is consistently low. The customer is a developer and the product is throughput.
• Best at: Latency. Tokens per second on their own silicon.
• Signals: No recent signal — NVIDIA disclosed a non-exclusive Groq technology license and the hiring of certain Groq employees in December 2025; the June 2026 financing shows Groq continuing as an operating company. Board state not certified.
• Tier: AI-native.
• Rubric: Company 7/12 — AI 3, Stage/equity 1, Design influence 1, Trajectory 2. Role 7/15 — Comp 2, Scope 1, Craft 2, AI exposure 1, Portfolio 1.
• Portfolio match: Thin, and I will not stretch it. The surface is a developer console. Model selection, rate limits, cost forecasting. Nothing published of yours addresses it, and the licensing-plus-personnel transfer clouds both the equity story and the question of who actually holds product authority.
• Verdict: Ignore-until-trigger. Trigger: a Head of Design or Director requisition clarifying independent product authority post-license. Until then it is a line item, not a target.

2. The mandate is funded. The location field is wrong.

Ten companies where a mandate you would want exists right now, advertised somewhere you cannot take it. Infuriating, and still the highest-yield cluster on the list, because the requisition is proof the mandate is funded. High confidence on current geography; every entry links to the primary source I read, either the live posting or, where the board itself is the constraint (Intercom, Monzo), the careers board. The location field is the only thing standing between you and a written, budgeted job. Save the searches on role titles and let them run.

1. Databricks. Unified data and AI platform on lakehouse architecture, now shipping applied AI products on top.

• Stage / Size / Capital: Late-stage private, very large.
• Offices: San Francisco HQ, but the relevant Staff Product Designer, AI Products role is listed in New York City.
• Value chain: Midstream data platform pushing downstream into applications. Design influence concentrates in the AI Products group.
• Best at: Putting enterprise data governance and model training inside one system.
• Signals: Qualifying Staff role live, New York only. Geographic mismatch is a hard gate.
• Tier: Enterprise platform.
• Rubric: Company 9/12 — AI 3, Stage/equity 2, Design influence 2, Trajectory 2. Role 12/15 — Comp 3, Scope 2, Craft 3, AI exposure 3, Portfolio 1.
• Portfolio match: AI Products is a trust-in-generated-queries problem: a data scientist reviewing SQL an agent wrote against tables they did not build. Brand Pulse's sentiment-attribution surface is the nearest published proof, machine-generated conclusions that had to show their sources before anyone would act on them. Differentiator: you have designed attribution for outputs nobody could verify at a glance, which is exactly this reviewer's position.
• Verdict: Ignore-until-trigger. Trigger: the same AI Products org opens a Staff+ requisition in San Francisco or US-remote.

2. Datadog. Cloud observability. Infrastructure monitoring, application performance, and log management in one platform.

• Stage / Size / Capital: Public.
• Offices: New York HQ. Senior Staff, Design Lab plus Staff roles for APM and Modern Compute, all New York.
• Value chain: Midstream operational tooling. Design owns the console where engineers decide whether an anomaly is an incident.
• Best at: Correlating a spike across metrics, traces, and logs fast enough to matter during an outage.
• Signals: Three distinct product mandates live, all geographically blocked. Read them as three territories, not one role posted three times.
• Tier: Enterprise platform.
• Rubric: Company 8/12 — AI 2, Stage/equity 2, Design influence 2, Trajectory 2. Role 12/15 — Comp 3, Scope 2, Craft 3, AI exposure 2, Portfolio 2.
• Portfolio match: Modern Compute is a control-surface problem. Show an engineer what an automated remediation is about to do to production, and give them an abort path. That is the Verify rung of your trust ladder, the state between watching a system and letting it act. Differentiator: you wrote the public framework for that intermediate state, and Datadog's users live in it all day.
• Verdict: Ignore-until-trigger. Trigger: any one of those three named mandates adds San Francisco or US-remote eligibility. Alert on role title, not company.

3. Miro. Visual collaboration canvas for distributed product and design teams, adding AI-assisted synthesis.

• Stage / Size / Capital: Late-stage private.
• Offices: Amsterdam HQ; the Director, Product Design role is advertised for Copenhagen, London, and New York.
• Value chain: Downstream collaboration application. Design influence is high by default, because the product is itself a design surface.
• Best at: Keeping an infinite canvas fast and legible with dozens of simultaneous editors.
• Signals: Director-level mandate live and geographically blocked; a Staff Prototypes search is also European-facing.
• Tier: Growth-stage platform.
• Rubric: Company 8/12 — AI 2, Stage/equity 2, Design influence 3, Trajectory 1. Role 12/15 — Comp 2, Scope 3, Craft 3, AI exposure 2, Portfolio 2.
• Portfolio match: Miro's live problem is AI synthesis on a canvas users already trust. The system clusters 200 sticky notes. Whose interpretation is that, and how do you disagree with it? Equinox+ is the structural proof: five brands into one MVP in three months, where the whole job was making a merged view feel authored instead of imposed. Differentiator: you have shipped consolidation work people adopted rather than resisted, at 4.8 stars.
• Verdict: Ignore-until-trigger. Trigger: a Staff+ or Director product design requisition naming San Francisco or US-remote.

4. Suno. Text-to-music generation for consumers and creators.

• Stage / Size / Capital: Well-funded private, AI-native.
• Offices: Cambridge, MA; the Senior/Staff Product Designer role is live in New York City.
• Value chain: Downstream consumer application over proprietary generative models. Design owns the entire perceived product.
• Best at: Compressing a complex generative pipeline into a prompt box non-musicians will actually use.
• Signals: Qualifying Senior/Staff role live in NYC. The earlier Head of Product Design search is stale. Do not cite it.
• Tier: AI-native.
• Rubric: Company 10/12 — AI 3, Stage/equity 2, Design influence 3, Trajectory 2. Role 11/15 — Comp 2, Scope 2, Craft 3, AI exposure 2, Portfolio 2.
• Portfolio match: Suno's unsolved surface is iterative revision. A user likes 80% of a generated track and has no vocabulary for the other 20%. Equinox+ is your proof of designing for people who know what they want and cannot name it. Differentiator: you have run consumer 0-to-1 on a three-month clock, which is this company's tempo.
• Verdict: Ignore-until-trigger. Company 10, the highest score in the cluster and the one to watch hardest. Trigger: the NYC requisition adds Bay Area or US-remote eligibility, or a separate SF Staff+ seat opens.

5. Headway. Marketplace connecting patients to in-network mental health providers and handling insurance billing.

• Stage / Size / Capital: Late-stage private.
• Offices: New York. Staff roles for Group Practices and Provider Onboarding, both NY.
• Value chain: Downstream healthcare marketplace between payers, providers, and patients. Value concentrates at the eligibility-and-billing decision.
• Best at: Making insurance actually work for therapy, which almost nobody has solved.
• Signals: Two distinct Staff mandates live, both geographically blocked.
• Tier: Healthcare/Regulated.
• Rubric: Company 9/12 — AI 2, Stage/equity 3, Design influence 2, Trajectory 2. Role 12/15 — Comp 2, Scope 3, Craft 3, AI exposure 2, Portfolio 2.
• Portfolio match: Provider Onboarding is multi-party credentialing. Provider, practice, payer, and platform each hold a piece of one truth. Thermo Fisher mySupply is the direct analogue: six pharma partners onto one system, 100% adoption in twelve months, because onboarding was designed as shared state instead of as a form. Differentiator: you have hit full adoption on multi-party onboarding in a regulated industry. That is the precise risk this panel is trying to retire.
• Verdict: Ignore-until-trigger. Trigger: either named mandate adds US-remote or Bay Area eligibility.

6. Duolingo. Gamified language learning with AI-driven lesson generation and conversational practice.

• Stage / Size / Capital: Public.
• Offices: Pittsburgh HQ and New York; Staff Product Designer roles live in both.
• Value chain: Downstream consumer application. Design and content are the product, and the influence ceiling is genuinely high.
• Best at: Retention mechanics. Nobody in consumer software is better at getting a user back tomorrow.
• Signals: Qualifying Staff roles live; no Bay Area or remote eligibility found.
• Tier: Growth-stage platform.
• Rubric: Company 8/12 — AI 2, Stage/equity 2, Design influence 3, Trajectory 1. Role 11/15 — Comp 2, Scope 2, Craft 3, AI exposure 2, Portfolio 2.
• Portfolio match: The AI tutor problem is confidence calibration for a learner who cannot evaluate the correction, including the case where the model says your sentence is wrong and it isn't. The Red Cross decision that transfers is showing the basis for guidance, so a user with no domain knowledge can decide how far to trust it. Differentiator: you have designed for people who must act on a system's word, a narrower skill than consumer craft generally.
• Verdict: Ignore-until-trigger. Trigger: a Staff+ product design role adds Bay Area or US-remote eligibility.

7. Intercom. Customer service platform with AI agents resolving support conversations end to end.

• Stage / Size / Capital: Private, profitable.
• Offices: Dublin and San Francisco, though current senior product design hiring on the careers board concentrates in London and Dublin.
• Value chain: Downstream service application. The AI agent resolves or escalates and design owns that boundary, which puts real decision rights in the design seat.
• Best at: Resolution rate. They publish it, which most competitors will not.
• Signals: No qualifying US-based senior design role found; current senior hiring is Europe-facing.
• Tier: AI-native.
• Rubric: Company 9/12 — AI 3, Stage/equity 2, Design influence 2, Trajectory 2. Role 12/15 — Comp 2, Scope 2, Craft 3, AI exposure 3, Portfolio 2.
• Portfolio match: The hardest surface here is the escalation seam. A customer four minutes into an agent conversation who now needs a human with no context. Your five-handoffs framework treats that as a distinct failure class, not a routing bug. Differentiator: you can bring a published taxonomy of handoff failures to a company whose economics depend on getting handoffs right.
• Verdict: Ignore-until-trigger. Trigger: a Staff or Principal product design role appears on the Intercom board with a US location.

8. Cohere. Enterprise-focused large language models emphasizing private and on-premise deployment.

• Stage / Size / Capital: Late-stage private.
• Offices: Toronto HQ; the current Senior Product Designer listing names New York, London, Paris, and Toronto.
• Value chain: Upstream model provider moving into enterprise applications. Design influence rises only in the application layer.
• Best at: Deploying capable models inside environments that cannot send data to a public API.
• Signals: Current listing sits below Staff level and outside the Bay Area. Fails both gates.
• Tier: AI-native.
• Rubric: Company 8/12 — AI 3, Stage/equity 2, Design influence 2, Trajectory 1. Role 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 3, Portfolio 1.
• Portfolio match: In a private-deployment context the design job is proving to a compliance officer that nothing left the perimeter, with governance visible in the product rather than in a PDF. Thermo Fisher's chain-of-custody surface is the proof, tied to $20M+ in margin because partners trusted what they could see. Differentiator: your governance work carries a commercial number, which is unusual and lands hard with enterprise buyers.
• Verdict: Ignore-until-trigger. Trigger: a Staff+ product design role explicitly including San Francisco or US-remote.

9. Mistral. Open-weight and commercial large language models plus an assistant product.

• Stage / Size / Capital: Late-stage private, European.
• Offices: Paris HQ, London; the current Product Designer listing is Paris/London.
• Value chain: Upstream model provider with a thin application layer. The design mandate is early and undefined.
• Best at: Shipping capable open-weight models enterprises can self-host.
• Signals: Current listing sits below the seniority threshold and outside the geographic gate.
• Tier: AI-native.
• Rubric: Company 7/12 — AI 3, Stage/equity 2, Design influence 1, Trajectory 1. Role 9/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 1.
• Portfolio match: None specific, and I will not manufacture one. The design mandate here is undefined, which means the portfolio conversation cannot be scoped until a trigger names a surface.
• Verdict: Ignore-until-trigger. Trigger: a Staff+ or design leadership requisition opens in a US office or US-remote.

10. Monzo. Digital bank with real-time spending controls and budgeting built into the account.

• Stage / Size / Capital: Late-stage private, UK.
• Offices: London, Cardiff, Dublin, Barcelona.
• Value chain: Downstream consumer banking. Design owns the primary relationship a customer has with their money.
• Best at: Turning a bank account into a product people open voluntarily.
• Signals: No qualifying US-based senior product design role; the careers footprint is entirely European.
• Tier: Healthcare/Regulated (high-consequence evaluation).
• Rubric: Company 8/12 — AI 1, Stage/equity 2, Design influence 3, Trajectory 2. Role 9/15 — Comp 1, Scope 2, Craft 3, AI exposure 1, Portfolio 2.
• Portfolio match: Monzo's fraud-intervention flow, freezing a card mid-transaction and asking the customer to confirm, is a reversibility surface for a stressed non-expert. No differentiation here. Plenty of consumer fintech designers tell that story, and with AI centrality at 1 the work would not strengthen your portfolio anyway.
• Verdict: Ignore-until-trigger. Low priority even after the gate opens. Trigger: first US-based Staff+ product design requisition.

3. Right company, wrong requisition

The design requisition exists. It has either aged past usefulness or been pitched below your altitude, and both produce identical monitoring behavior: watch the board for the next requisition ID, not for the company to change. High confidence on every posting state below; each was checked against its primary source and linked where the page loaded. Forty-three days is my decay threshold, not a law of nature, and LinkedIn reports an average of 66 days to make a hire. A 54-day-old requisition is not necessarily dead. It is one where a full slate assembled seven weeks ago and you would be arguing against an incumbent finalist. That is why these are Ignore and not Watch.

1. OpenAI. Frontier models plus ChatGPT and a developer platform.

• Stage / Size / Capital: Late-stage private, very large.
• Offices: San Francisco.
• Value chain: Vertically integrated across model, application, and platform. Value concentrates everywhere, which paradoxically diffuses design authority.
• Best at: Consumer distribution of a frontier model. No competitor is close.
• Signals: Product Design Manager and Product Designer listings present on the careers search; no date-certified Director+ or Staff/Senior Staff mandate established this pass.
• Tier: AI-native.
• Rubric: Company 10/12 — AI 3, Stage/equity 2, Design influence 2, Trajectory 3. Role 12/15 — Comp 3, Scope 2, Craft 3, AI exposure 3, Portfolio 1.
• Portfolio match: The live problem is agent supervision inside ChatGPT. A consumer watches a task run and decides whether to intervene. Your trust ladder was written for that progression, and TinyFish gives you present-tense evidence of running it against real enterprise agent traces. Differentiator: you published the framework before the market needed it, which reads as conviction rather than opportunism.
• Verdict: Ignore-until-trigger. Trigger: a Staff/Senior Staff or Director+ requisition with a direct role URL and a fresh posting date. Check the careers search directly. Aggregators miss this company.

2. Figma. Collaborative interface design and prototyping, expanding into slides, sites, and code generation.

• Stage / Size / Capital: Public since July 2025.
• Offices: San Francisco.
• Value chain: Downstream creative tooling. Design influence is definitionally maximal.
• Best at: Multiplayer editing at a fidelity nobody has replicated.
• Signals: No recent signal — the Greenhouse board carries Manager, Product Design and Product Designer roles, no qualifying Staff+ or Director mandate in the checked set (verified Aug 1).
• Tier: Enterprise platform.
• Rubric: Company 8/12 — AI 2, Stage/equity 2, Design influence 3, Trajectory 1. Role 11/15 — Comp 3, Scope 2, Craft 3, AI exposure 2, Portfolio 1.
• Portfolio match: Figma's generative surfaces face a credibility problem specific to their users. Designers notice every wrong choice the model makes. That is the Verify rung inverted, with the user more expert than the system. Nothing in your published work addresses expert-over-machine review. Flag it as a real gap rather than papering over it, because this panel will find it.
• Verdict: Ignore-until-trigger. Trigger: a requisition crossing the Staff or Director threshold on the Greenhouse board.

3. Zoox. Purpose-built autonomous robotaxi with no steering wheel, operating in Las Vegas and the Bay Area.

• Stage / Size / Capital: Amazon subsidiary.
• Offices: Foster City.
• Value chain: Vertically integrated autonomy across vehicle, fleet, and rider app. Corporate ownership caps the equity thesis.
• Best at: Designing the vehicle around the rider instead of retrofitting autonomy into a car.
• Signals: Senior Manager, Product and Experience Design is live on the Lever board; no Director+, Staff, or Senior Staff product design role in the checked set.
• Tier: Healthcare/Regulated (high-consequence evaluation).
• Rubric: Company 8/12 — AI 3, Stage/equity 1, Design influence 2, Trajectory 2. Role 11/15 — Comp 2, Scope 2, Craft 2, AI exposure 3, Portfolio 2.
• Portfolio match: Rider assistance, meaning what a passenger can request when the vehicle stops for reasons they cannot see, is a delegation-failure problem with nobody in the seat. Carrier IQ's escalation path is the structural analogue: automation running unattended with a defined human recourse. Differentiator: real on mechanism, thin on outcome. Amazon ownership keeps the equity case weak no matter how well the fit reads.
• Verdict: Ignore-until-trigger. Scope play only. Trigger: a Director+ or Staff IC product design requisition on the Zoox Lever board.

4. Together AI. GPU cloud and inference platform for open-source models, plus fine-tuning tooling.

• Stage / Size / Capital: Late-stage private.
• Offices: San Francisco.
• Value chain: Upstream compute and inference. Design influence limited by a developer-primary audience.
• Best at: Running open models cheaply at scale behind a usable API.
• Signals: Lead Product Designer live in San Francisco on the Greenhouse board, below the qualifying title threshold.
• Tier: AI-native.
• Rubric: Company 7/12 — AI 3, Stage/equity 2, Design influence 1, Trajectory 1. Role 9/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 1.
• Portfolio match: A Lead posting at this size usually means the design function is being founded, not filled. If it converts upward, the story is function-building, and Alibaba is your proof: you named a structural gap in the North America org, built the mandate, and ran three cross-functional sprints against it. Differentiator: you have created a design mandate that did not previously exist, which is the actual job here.
• Verdict: Ignore-until-trigger. Trigger: the same org opens Head/Director or Staff+ product design. Lead-to-Head conversion inside 90 days is common enough to justify a monitor. Pattern-inferred, not validated.

5. Palantir. Data integration and operational decision platforms for government and large enterprise.

• Stage / Size / Capital: Public.
• Offices: Denver HQ, Palo Alto, Seattle, Washington DC.
• Value chain: Vertically integrated from ingestion through decision application. Design owns consequential surfaces here in a way it rarely does at this scale.
• Best at: Putting heterogeneous data behind a single operational decision in environments where being wrong has physical consequences.
• Signals: A Product Designer posting is live; no qualifying Staff+/Director title established this pass.
• Tier: Enterprise platform.
• Rubric: Company 9/12 — AI 3, Stage/equity 1, Design influence 3, Trajectory 2. Role 12/15 — Comp 2, Scope 2, Craft 3, AI exposure 3, Portfolio 2.
• Portfolio match: AIP surfaces put an analyst in the position of approving a machine-recommended action with real-world consequences and a legal record. Carrier IQ is your published instance: automated insurance quoting where a human's approval carried liability. Differentiator: you have designed the accountability moment itself, not the model and not the dashboard wrapped around it.
• Verdict: Ignore-until-trigger. Trigger: a Staff, Senior Staff, or Director product design requisition under a new ID.

6. Glean. Enterprise search and assistant across a company's internal applications.

• Stage / Size / Capital: Late-stage private.
• Offices: Palo Alto; the current Product Designer listing is in Bangalore.
• Value chain: Midstream knowledge layer over enterprise systems. Design owns whether employees believe the answers.
• Best at: Permission-aware retrieval, surfacing only what the asking employee is cleared to see.
• Signals: Current listing fails both seniority and geography gates.
• Tier: Enterprise platform.
• Rubric: Company 9/12 — AI 3, Stage/equity 3, Design influence 2, Trajectory 1. Role 11/15 — Comp 2, Scope 2, Craft 2, AI exposure 3, Portfolio 2.
• Portfolio match: Glean's core surface is attribution under permission constraints. Here is the answer, here is the source, and no, you cannot open it. Your Alibaba product-detail work transfers: trust signals assembled from sources of varying reliability so a buyer could act, worth +2.2 points of NPS on a platform where nobody trusted anything. Differentiator: you have moved a measured trust metric on a marketplace, which is harder than moving a task-completion metric.
• Verdict: Ignore-until-trigger. Trigger: a Staff+ or Director requisition in Palo Alto, San Francisco, or US-remote.

7. Ramp. Corporate cards, bill pay, and procurement with automated expense controls.

• Stage / Size / Capital: Late-stage private, fast-growing.
• Offices: New York, San Francisco.
• Value chain: Downstream finance application with strong AI automation. High ceiling, and the function is led by a VP Design.
• Best at: Removing finance-team labor instead of reorganizing it.
• Signals: No recent signal — Director, Product Design first published June 8, 2026, 54 days old at cutoff, past the 43-day threshold. The page still loads, which is not the same as active.
• Tier: Growth-stage platform.
• Rubric: Company 10/12 — AI 3, Stage/equity 3, Design influence 2, Trajectory 2. Role 13/15 — Comp 3, Scope 3, Craft 3, AI exposure 3, Portfolio 1.
• Portfolio match: Ramp's live problem is agentic procurement. Software that negotiates and buys on a company's behalf, with a controller who has to be able to reverse it. Retail Velocity is your analogue for machine-initiated action reaching someone with authority to stop it. Differentiator: you have shipped an autonomous-recommendation surface with a defined stop path, and you are running the enterprise version of that problem in production right now.
• Verdict: Ignore-until-trigger. Highest combined score in the cluster and the most painful entry on the list. Trigger: a new Ashby role ID, an authoritative repost date, or Diego Zaks reopening the search publicly. The live Head of Brand posting reports to VP Design, which confirms the seat is filled and reporting into it.

8. Rubrik. Data security and cyber-recovery. Backup, ransomware detection, and restore orchestration.

• Stage / Size / Capital: Public.
• Offices: Palo Alto.
• Value chain: Midstream data protection. Design owns the recovery decision, which is genuinely high-stakes.
• Best at: Getting an enterprise back online after ransomware without restoring the ransomware.
• Signals: No recent signal — the Director/Senior Manager Product Design requisition was first recorded February 5, 2026; a July 31 board update did not establish a reopened search, and the title has rendered inconsistently between Director and Senior Manager.
• Tier: Enterprise platform.
• Rubric: Company 7/12 — AI 2, Stage/equity 1, Design influence 2, Trajectory 2. Role 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 2.
• Portfolio match: The restore decision under active compromise is a time-critical judgment by someone who cannot verify the system's recommendation. The transferable Red Cross decision is the consolidation itself: six systems collapsed into one so a single operator could act without assembling context first. Differentiator: you have designed for crisis-mode operators specifically, not merely for complexity.
• Verdict: Ignore-until-trigger. Read the title inconsistency as a signal. A Director requisition that renders as Senior Manager usually means the level was never settled internally, which predicts a scope fight after you join. Trigger: a new Greenhouse requisition ID with a stable title.

9. Salesforce. Enterprise CRM and the Agentforce agent platform.

• Stage / Size / Capital: Public, very large.
• Offices: San Francisco.
• Value chain: Vertically integrated enterprise platform. Design influence varies enormously by business unit, and that variance is the central risk.
• Best at: Distribution. Agentforce reaches more enterprise seats than any competing agent platform.
• Signals: No recent signal — VP, Product Design, Informatica posted May 15, 2026; 78 days old at cutoff.
• Tier: Enterprise platform.
• Rubric: Company 8/12 — AI 3, Stage/equity 1, Design influence 2, Trajectory 2. Role 11/15 — Comp 3, Scope 2, Craft 1, AI exposure 3, Portfolio 2.
• Portfolio match: The post-acquisition Informatica mandate is about making data governance legible inside agent workflows, as reusable patterns a hundred product teams will inherit. Alibaba is the scale proof: patterns across homepage, search, and product detail that other teams adopted, on a platform moving $50B+ in GMV. Differentiator: you have set reusable pattern standards at that scale. That is what a VP-level Salesforce panel probes for.
• Verdict: Ignore-until-trigger. Note the honest tension in the scores. Craft depth is 1. This is a governance and org-design job. Trigger: a new Informatica design leadership requisition under Salesforce UX with a new job ID.

10. Adobe. Creative and document software with generative AI woven through Firefly and Acrobat.

• Stage / Size / Capital: Public, very large.
• Offices: San Jose, San Francisco.
• Value chain: Downstream creative and document applications. Mature design org led by a VP of Design.
• Best at: Shipping generative features into workflows professionals already depend on, without breaking them.
• Signals: No recent signal — prior Director-level design requisitions were marked filled in the August 1 check while the broader careers site stayed active. Board-level URL only; the specific requisition pages no longer resolve.
• Tier: Enterprise platform.
• Rubric: Company 8/12 — AI 2, Stage/equity 1, Design influence 3, Trajectory 2. Role 10/15 — Comp 3, Scope 2, Craft 2, AI exposure 2, Portfolio 1.
• Portfolio match: Adobe's live tension is provenance in generated assets. Content credentials that survive editing and export, which is a governance artifact living inside a creative tool. Thermo Fisher's chain-of-custody design is the transferable pattern. Differentiator: honestly modest. Adobe has deep internal craft. Your edge is the governance layer, not the creative surface.
• Verdict: Ignore-until-trigger. Trigger: a new qualifying requisition number, not a recycled status page. Eric Snowden is VP of Design and the named buyer if one appears.

4. No design seat exists yet

The forming-org pattern. Serious AI products, no visible product design leadership, so the first senior design hire defines the function instead of inheriting it. Highest upside on the list and highest risk, because a company that has gone this long without a design leader has usually built decision habits that route around design. Mechanism reliability: pattern-inferred. I have no validated lead time from "AI company reaches scale" to "first design leadership requisition," and my honest read is that the trigger is not time-based at all. Watch for a design-adjacent role appearing: design engineer, design systems, design ops. That is the tell that somebody inside finally owns the argument. One gap I will name rather than fill. I could not source a named head of design engineering or design systems at Sierra, Perplexity, or Writer this pass, so at those three the monitor is the board, not a person.

1. Anthropic. Claude models plus consumer, developer, and enterprise products.

• Stage / Size / Capital: Late-stage private, very large.
• Offices: San Francisco.
• Value chain: Vertically integrated from frontier model through application. Value concentrates at the model; the enterprise surface is where design would live.
• Best at: Enterprise trust. They win deals on safety posture rather than benchmark scores.
• Signals: No qualifying senior product design role on the official careers board (verified Aug 1). Absence of visible product design leadership at this scale is itself the signal.
• Tier: AI-native.
• Rubric: Company 11/12 — AI 3, Stage/equity 3, Design influence 2, Trajectory 3. Role 13/15 — Comp 3, Scope 3, Craft 3, AI exposure 3, Portfolio 1.
• Portfolio match: The unowned surface is model evidence: showing an enterprise user why Claude reached a conclusion, in a form that survives an internal audit. Your five-handoffs framework is the only published model I know that treats evidence presentation as a discrete design object rather than a transparency feature. Differentiator: you wrote it before this company needed to solve it, and you test it against production agent traces at TinyFish daily.
• Verdict: Ignore-until-trigger. Highest company score on the list and completely unactionable. Trigger: any design-adjacent requisition on the Anthropic board. That is the tell. Not the leadership role.

2. Sierra. Conversational AI agents for customer service, deployed per-company with brand-specific behavior.

• Stage / Size / Capital: Late-stage private, AI-native.
• Offices: San Francisco.
• Value chain: Downstream service application. The agent resolves, refunds, and escalates, which puts extraordinary decision weight on design.
• Best at: Constraining agent behavior to a company's actual policies instead of a generic persona.
• Signals: No qualifying senior product design role on the Ashby board (verified Aug 1).
• Tier: AI-native.
• Rubric: Company 10/12 — AI 3, Stage/equity 3, Design influence 2, Trajectory 2. Role 13/15 — Comp 3, Scope 3, Craft 3, AI exposure 3, Portfolio 1.
• Portfolio match: Sierra's least-solved surface is the operator console, where a customer-service manager watches agent conversations and decides which behaviors to correct. Supervision of an autonomous system by a non-engineer. Retail Velocity is your published instance: field operators reviewing machine-generated priorities and overriding them. Differentiator: you have designed the supervisor's view, which almost every agent company builds last and badly.
• Verdict: Ignore-until-trigger. Trigger: any product design requisition at any level on the Sierra Ashby board. At this stage the first one is the one.

3. Perplexity. AI answer engine with citations, plus a browser and shopping surfaces.

• Stage / Size / Capital: Late-stage private.
• Offices: San Francisco.
• Value chain: Downstream consumer application over retrieval and third-party models. Design owns the entire trust proposition.
• Best at: Making citations feel native to an answer instead of bolted onto it.
• Signals: No qualifying senior product design role on the Ashby board (verified Aug 1).
• Tier: AI-native.
• Rubric: Company 10/12 — AI 3, Stage/equity 3, Design influence 3, Trajectory 1. Role 12/15 — Comp 2, Scope 3, Craft 3, AI exposure 3, Portfolio 1.
• Portfolio match: Answer provenance under commercial pressure. As shopping surfaces monetize, citation credibility and merchant incentive collide. You have run that exact tension: Alibaba's product-detail redesign had to make trust signals credible on a platform where sellers paid for placement, and NPS moved 2.2 points anyway. Differentiator: you have defended trust design against a monetization roadmap and won on the metrics, which is a political skill as much as a craft one.
• Verdict: Ignore-until-trigger. Trigger: a design systems or design ops requisition. That formalizing signal has historically preceded a product design leadership search, on a small sample. Monitor it. Do not forecast from it.

4. Character AI. Consumer platform for conversational AI characters and companions.

• Stage / Size / Capital: Private; a structured licensing arrangement with Google constrains the independent equity story.
• Offices: Menlo Park.
• Value chain: Downstream consumer application. Design owns persona presentation and safety framing.
• Best at: Retention on open-ended conversation, unmatched at consumer scale.
• Signals: No qualifying senior product design role on the Ashby board (verified Aug 1).
• Tier: AI-native.
• Rubric: Company 7/12 — AI 3, Stage/equity 1, Design influence 2, Trajectory 1. Role 9/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 1.
• Portfolio match: Persona continuity and safety disclosure to vulnerable users, many of them young, is non-expert trust calibration carrying real duty-of-care weight. Differentiator: real on the audience. But the licensing structure caps equity and the safety exposure is significant. I would not prioritize this.
• Verdict: Ignore-until-trigger. Trigger: a design leadership requisition plus clarified employee equity terms. Both. Not either.

5. Writer. Enterprise generative AI platform with company-specific models and governed content workflows.

• Stage / Size / Capital: Late-stage private.
• Offices: San Francisco, New York.
• Value chain: Midstream enterprise application layer. Design owns the governance surface, which is where the differentiation actually lives.
• Best at: Enterprise content governance, with brand and compliance rules enforced at generation time.
• Signals: No qualifying senior product design role on the Ashby board (verified Aug 1).
• Tier: Enterprise platform.
• Rubric: Company 9/12 — AI 3, Stage/equity 3, Design influence 2, Trajectory 1. Role 12/15 — Comp 2, Scope 3, Craft 3, AI exposure 3, Portfolio 1.
• Portfolio match: The review surface for regulated marketing content, where legal and medical reviewers approve AI-generated copy under sign-off liability, maps straight to Carrier IQ's approval gate. Differentiator: you have designed an approval gate inside an automated flow, which is the surface Writer's enterprise customers route around when it is built badly.
• Verdict: Ignore-until-trigger. Scores better than Grammarly on the same mechanism. If both go live, take this one. Trigger: first Staff+ or Head of Design requisition on the Writer Ashby board.

6. ElevenLabs. Voice synthesis, dubbing, and conversational voice agents.

• Stage / Size / Capital: Late-stage private, AI-native.
• Offices: New York, London; distributed.
• Value chain: Midstream generative media infrastructure with a growing application layer.
• Best at: Voice quality at latency low enough for real-time conversation.
• Signals: No qualifying senior product design role in the reserve board pass (verified Aug 1).
• Tier: AI-native.
• Rubric: Company 8/12 — AI 3, Stage/equity 3, Design influence 1, Trajectory 1. Role 10/15 — Comp 2, Scope 2, Craft 3, AI exposure 2, Portfolio 1.
• Portfolio match: Voice-agent design has almost no visual surface, so the trust work lives in turn-taking, interruption, and disclosure. Genuinely a different medium from anything you have published. No differentiation here. Your framework transfers conceptually. Your artifacts do not.
• Verdict: Ignore-until-trigger. Verify geography before spending time. Trigger: a US-eligible Staff+ or Head of Design requisition.

7. Replit. Browser-based development environment with an AI agent that builds and deploys applications.

• Stage / Size / Capital: Late-stage private.
• Offices: Foster City.
• Value chain: Downstream developer application over cloud infrastructure. The agent writes and ships code, which is about as much decision weight as a product can carry.
• Best at: Taking a non-engineer from prompt to deployed app without leaving the browser.
• Signals: No qualifying senior product design role in the reserve board pass (verified Aug 1).
• Tier: AI-native.
• Rubric: Company 9/12 — AI 3, Stage/equity 3, Design influence 2, Trajectory 1. Role 12/15 — Comp 2, Scope 3, Craft 3, AI exposure 3, Portfolio 1.
• Portfolio match: Sharpest unsolved surface is rollback. A non-engineer whose agent broke a working app, with no mental model of version control. Reversibility for a user who skipped the first two rungs of the trust ladder entirely. Differentiator: TinyFish gives you the engineering-adjacent version of this daily, so you can describe what agent failure looks like in production rather than in a demo.
• Verdict: Ignore-until-trigger. Trigger: first Staff+ product design requisition.

8. LangChain. Open-source framework and hosted tooling for building and observing LLM applications.

• Stage / Size / Capital: Private, developer tooling.
• Offices: San Francisco.
• Value chain: Upstream developer framework. Design influence structurally low; the primary artifact is a library.
• Best at: Trace-level observability for agent runs. LangSmith is the strongest product in the portfolio.
• Signals: No qualifying senior product design role in the reserve board pass (verified Aug 1).
• Tier: AI-native.
• Rubric: Company 7/12 — AI 3, Stage/equity 2, Design influence 1, Trajectory 1. Role 9/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 1.
• Portfolio match: Agent-trace inspection, reading why an agent chose a tool three steps back, is the one place design matters here. It is also your daily reality at TinyFish. Current-role currency, not published proof, and this company will want proof. Weak on the portfolio side.
• Verdict: Ignore-until-trigger. Trigger: a Head of Design requisition, not an IC one. An IC role at this company reports into engineering.

9. Anrok. Sales tax compliance automation for software companies.

• Stage / Size / Capital: Growth-stage private.
• Offices: San Francisco.
• Value chain: Midstream compliance layer between billing systems and tax authorities. Narrow, high-consequence.
• Best at: Determining tax obligation across jurisdictions automatically, which is genuinely hard and genuinely thankless.
• Signals: No qualifying senior product design role on the Ashby board (verified Aug 1).
• Tier: Healthcare/Regulated (high-consequence evaluation).
• Rubric: Company 6/12 — AI 1, Stage/equity 2, Design influence 2, Trajectory 1. Role 9/15 — Comp 2, Scope 2, Craft 2, AI exposure 1, Portfolio 2.
• Portfolio match: A finance lead accepts an automated tax determination they cannot independently verify and will personally answer for. Structurally identical to Carrier IQ, lower stakes. Differentiator: strong positioning fit, weak company score. AI centrality is the binding constraint here, not your story.
• Verdict: Ignore-until-trigger. Trigger: a Head of Design or Staff requisition combined with evidence AI has moved from feature to core in the product.

10. Pinecone. Managed vector database for retrieval in AI applications.

• Stage / Size / Capital: Late-stage private.
• Offices: New York, San Francisco.
• Value chain: Upstream retrieval infrastructure. The layer where design influence is most reliably capped: the buyer is an ML engineer and the product is an index.
• Best at: Retrieval quality at scale without managing infrastructure.
• Signals: Unverified — direct board state not certified this pass. Do not read that as absence.
• Tier: AI-native.
• Rubric: Company 7/12 — AI 3, Stage/equity 2, Design influence 1, Trajectory 1. Role 8/15 — Comp 2, Scope 1, Craft 2, AI exposure 2, Portfolio 1.
• Portfolio match: The only interesting surface is retrieval debugging, showing an engineer why the wrong chunk came back. Real design work, small audience, low org influence. The value-chain rule applies cleanly and argues against.
• Verdict: Ignore-until-trigger. Structurally low-ceiling for a design leader. Trigger: a Head/Director design requisition naming an application surface rather than a console.

5. The seat is filled and the board is quiet

Mature organizations with a leader of record, or a product executive who has been in seat long enough to have posted and hasn't. Two things reopen these: a departure or a mandate expansion. Confidence on the seat identifications runs high to moderate, because several come from current conference biographies and leadership pages rather than from the companies themselves, which means a change can be weeks old before it surfaces anywhere. I found no validated lead time between a design leader leaving and a successor requisition appearing, so treat the departure as the alert and the requisition as the action. Where I could not source a leader of record (Vercel, Cloudflare, Elastic, Roblox, Motive, Airbnb) that is a research gap, not evidence of a vacancy, and the monitor is the board.

1. Notion. Connected workspace for docs, databases, and projects, with AI search and agents layered in.

• Stage / Size / Capital: Late-stage private.
• Offices: San Francisco.
• Value chain: Downstream productivity application. Design influence is high and culturally embedded.
• Best at: A composable data model ordinary users can assemble.
• Signals: No recent signal — Randy Hunt joined as Head of Design May 16, 2024 and a 2026 conference biography still puts him at the head of the global design organization; no qualifying role on the Ashby board as of Aug 1.
• Tier: Growth-stage platform.
• Rubric: Company 8/12 — AI 2, Stage/equity 2, Design influence 3, Trajectory 1. Role 11/15 — Comp 2, Scope 2, Craft 3, AI exposure 2, Portfolio 2.
• Portfolio match: Notion's agents act inside a workspace where one wrong edit destroys somebody's system, which makes scoped autonomy in a shared document the design problem. The Delegate rung applied to a multi-tenant artifact. Differentiator: you hold a public position on when delegation is appropriate, and this design org would engage it on peer terms rather than as an interview exercise.
• Verdict: Ignore-until-trigger. Trigger: a Director or Staff product design requisition under Hunt, or a public change to his seat.

2. Airtable. Relational database as a collaborative app builder, with AI agents over structured data.

• Stage / Size / Capital: Late-stage private.
• Offices: San Francisco.
• Value chain: Downstream application platform. Design owns the abstraction that makes databases usable.
• Best at: Letting non-technical teams build operational tools that survive contact with reality.
• Signals: No recent signal — a current conference biography identifies Jaime McFarland as VP/Head of Design; no qualifying role on the Greenhouse board as of Aug 1.
• Tier: Growth-stage platform.
• Rubric: Company 8/12 — AI 2, Stage/equity 2, Design influence 3, Trajectory 1. Role 10/15 — Comp 2, Scope 2, Craft 3, AI exposure 2, Portfolio 1.
• Portfolio match: Airtable's agents write to structured records, so bulk-change preview is the design problem: 400 pending row edits shown in a way a human can approve or reject meaningfully. Retail Velocity is the volume analogue, and the decision that transfers is grouping rather than paging. Differentiator: you have solved review-at-volume, not just review.
• Verdict: Ignore-until-trigger. Trigger: a Staff+ or Director requisition on the Greenhouse board, or McFarland surfacing a search publicly.

3. Vercel. Frontend deployment platform with AI-assisted generation through v0.

• Stage / Size / Capital: Late-stage private.
• Offices: San Francisco.
• Value chain: Midstream developer platform pushing downstream through v0. Design influence rises with the v0 surface.
• Best at: Deployment ergonomics. The gap between commit and live is the product.
• Signals: No qualifying senior product design role on the Greenhouse board (verified Aug 1). No leader of record sourced this pass.
• Tier: AI-native.
• Rubric: Company 8/12 — AI 2, Stage/equity 2, Design influence 2, Trajectory 2. Role 11/15 — Comp 2, Scope 2, Craft 3, AI exposure 2, Portfolio 2.
• Portfolio match: v0's problem is iteration legibility. The user wants to change one thing about generated output and cannot say which thing. No differentiation here. Vercel's internal design bar is high, developer-experience craft is abundant in their candidate pool, and your genuine edge in agent trust is not what this surface is asking for yet.
• Verdict: Ignore-until-trigger. Trigger: a Staff+ or Director requisition naming v0 specifically, not platform design generally.

4. Snowflake. Cloud data warehouse expanding into AI application development.

• Stage / Size / Capital: Public.
• Offices: Bozeman HQ, San Mateo.
• Value chain: Midstream data platform. Design influence has historically been constrained by an engineering-led product culture.
• Best at: Separating storage from compute in a way enterprises can reason about financially.
• Signals: Zack Gottlieb stated publicly in early 2026 that he wrapped a two-year tenure as Head of Design. Self-reported, board state not certified this pass. Treat the seat as open, not the role as absent.
• Tier: Enterprise platform.
• Rubric: Company 7/12 — AI 2, Stage/equity 1, Design influence 2, Trajectory 2. Role 10/15 — Comp 3, Scope 2, Craft 2, AI exposure 2, Portfolio 1.
• Portfolio match: Snowflake's AI surfaces put a data engineer in the position of approving generated transformations against governed enterprise data. Thermo Fisher again, and the reason it fits here is that the job is making governance a normal step instead of a gate people resent. That is what drove 100% partner adoption. Differentiator: you have made compliance workflows adopted rather than tolerated, with a margin number attached.
• Verdict: Ignore-until-trigger. The only clean design-leadership departure on the recognized board and the highest-urgency mechanism anywhere here. Trigger: a successor announcement or a Head/VP/Director product design requisition. Set the alert now.

5. MongoDB. Document database with a managed cloud service, expanding into AI application infrastructure.

• Stage / Size / Capital: Public.
• Offices: New York HQ with distributed engineering.
• Value chain: Upstream data infrastructure with an emerging application layer. Design influence historically low.
• Best at: Developer adoption. The on-ramp from local database to managed cluster is the moat.
• Signals: No recent signal — MongoDB named Pablo Stern co-CPO for AI and emerging products and Ben Cefalo co-CPO for the core database business on April 22, 2026. 101 days before cutoff. Executive window closed, no qualifying design role on the board.
• Tier: Enterprise platform.
• Rubric: Company 6/12 — AI 2, Stage/equity 1, Design influence 1, Trajectory 2. Role 9/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 1.
• Portfolio match: Weak, and the org structure is why. A dual-CPO arrangement with a separate AI organization reads as design-influence risk, not opportunity. Two product executives usually means design reports two levels below anybody who decides. This is the only recognized target that fits the aged-executive pattern cleanly, and the pattern cannot lift a company score of 6.
• Verdict: Ignore-until-trigger. Trigger: a Staff/Senior Staff or Director requisition explicitly under Stern's AI and emerging products organization. Under Cefalo's core database org, pass.

6. Cloudflare. Edge network, security, and developer platform with AI inference at the edge.

• Stage / Size / Capital: Public.
• Offices: San Francisco, Austin.
• Value chain: Upstream network infrastructure with a developer platform layer. Design influence caps at the dashboard.
• Best at: Shipping infrastructure features at a cadence that embarrasses competitors.
• Signals: No qualifying senior product design role in the reserve board pass (verified Aug 1). No leader of record sourced this pass.
• Tier: Enterprise platform.
• Rubric: Company 7/12 — AI 2, Stage/equity 1, Design influence 2, Trajectory 2. Role 9/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 1.
• Portfolio match: Bot and agent management, deciding which automated traffic gets through, is a policy decision rendered as a dashboard. Genuinely adjacent to your TinyFish work, with no published proof to anchor it.
• Verdict: Ignore-until-trigger. Low priority. Trigger: a Director+ design requisition naming the AI or agent platform.

7. Elastic. Search and observability platform built on Elasticsearch, with vector search for AI applications.

• Stage / Size / Capital: Public.
• Offices: Distributed; US presence including Mountain View.
• Value chain: Midstream search infrastructure. Design owns the query and analysis surface.
• Best at: One engine serving search, logging, and security analytics instead of three systems.
• Signals: No qualifying senior product design role in the reserve board pass (verified Aug 1). No leader of record sourced this pass.
• Tier: Enterprise platform.
• Rubric: Company 7/12 — AI 2, Stage/equity 1, Design influence 2, Trajectory 2. Role 9/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 1.
• Portfolio match: Relevance tuning as a user-facing surface, letting a search administrator see and adjust why results rank as they do. Your Alibaba search sprint is the direct proof. Differentiator: you have tuned relevance against a commercial metric rather than a precision score.
• Verdict: Ignore-until-trigger. Trigger: a Staff+ requisition naming search relevance or vector search product surfaces.

8. Roblox. User-generated 3D experience platform with a creator economy and generative creation tools.

• Stage / Size / Capital: Public.
• Offices: San Mateo.
• Value chain: Vertically integrated across engine, marketplace, and client. Design influence is significant in the creator surfaces.
• Best at: Creator monetization at a scale no competing user-generated platform approaches.
• Signals: No qualifying senior product design role in the reserve board pass (verified Aug 1). No leader of record sourced this pass.
• Tier: Enterprise platform.
• Rubric: Company 8/12 — AI 2, Stage/equity 1, Design influence 3, Trajectory 2. Role 10/15 — Comp 2, Scope 2, Craft 3, AI exposure 2, Portfolio 1.
• Portfolio match: Generative creation tools for young creators combine safety and attribution. Red Cross covers the vulnerable-user framing, Brand Pulse covers attribution mechanics, and neither stands alone. Composite matches are weaker matches. Say so if you ever pitch here.
• Verdict: Ignore-until-trigger. Trigger: a Director+ requisition in creator tools or safety specifically.

9. Motive. Fleet management with dashcams, telematics, and AI safety scoring for commercial vehicles.

• Stage / Size / Capital: Late-stage private.
• Offices: San Francisco.
• Value chain: Vertically integrated hardware plus software for physical operations. Design owns the fleet manager's decision surface.
• Best at: Turning dashcam footage into coaching drivers accept rather than resent.
• Signals: No qualifying senior product design role on the Lever board (verified Aug 1). No leader of record sourced this pass.
• Tier: Enterprise platform.
• Rubric: Company 10/12 — AI 3, Stage/equity 3, Design influence 2, Trajectory 2. Role 12/15 — Comp 2, Scope 2, Craft 3, AI exposure 3, Portfolio 2.
• Portfolio match: AI safety scoring that affects a driver's employment is a consequential-judgment surface with a contestation requirement. The driver has to be able to dispute the machine. Carrier IQ's review gate is the mechanism, and what makes it fit is that both involve a machine judgment about a person with money attached. Differentiator: you have designed the appeal path, which most AI scoring products treat as a support ticket.
• Verdict: Ignore-until-trigger. Company 10, role 12. Corrected, that puts Motive among the top-scoring entries on the entire list and easily the quietest. Trigger: any Staff+ or Director product design requisition on the Lever board.

10. Airbnb. Marketplace for stays and experiences with AI-assisted search and support.

• Stage / Size / Capital: Public.
• Offices: San Francisco.
• Value chain: Downstream two-sided marketplace. Design influence is famously high and famously centralized.
• Best at: Design-led product decision-making at public-company scale, which almost nobody sustains.
• Signals: No qualifying senior product design role on the careers site (verified Aug 1). No leader of record sourced this pass.
• Tier: Enterprise platform.
• Rubric: Company 8/12 — AI 2, Stage/equity 1, Design influence 3, Trajectory 2. Role 11/15 — Comp 3, Scope 2, Craft 3, AI exposure 2, Portfolio 1.
• Portfolio match: The two-sided AI problem is asymmetric trust. A guest trusting an AI recommendation about a host's property; a host trusting an AI decision about their listing. Allē is your dual-surface proof, consumer and provider on one system at 3.2x redemption and $42 CAC, because each side believed different things about the same transaction. Differentiator: you have shipped a two-sided trust surface with measured economics on both sides.
• Verdict: Ignore-until-trigger. Craft and portfolio play, not a wealth play. Trigger: a Staff+ requisition in the AI or search organization.

6. AI is still a feature here

These clear the rubric on stage, scope, and compensation, and fail it on AI centrality. That condition fixes itself in exactly one observable way: a product launch that moves AI out of a sidebar and onto a decision surface. Mechanism real, lead time speculative. The launch is your outreach hook whether or not a requisition exists, because a company that just shipped an agent into a regulated workflow has a design problem it has not solved and knows it.

1. Carta. Cap table management, valuations, and private-market fund administration.

• Stage / Size / Capital: Late-stage private.
• Offices: San Francisco, New York.
• Value chain: Downstream financial infrastructure for private markets. Design owns the surface where equity decisions get recorded.
• Best at: Being the system of record everyone else has to integrate with.
• Signals: No qualifying senior product design role on the Greenhouse board (verified Aug 1).
• Tier: Healthcare/Regulated (high-consequence evaluation).
• Rubric: Company 5/12 — AI 1, Stage/equity 2, Design influence 1, Trajectory 1. Role 8/15 — Comp 2, Scope 2, Craft 2, AI exposure 1, Portfolio 1.
• Portfolio match: If Carta ships automated 409A valuation or waterfall modeling, the problem becomes a CFO signing off on a machine-generated number with audit consequences. Carrier IQ maps. The trigger has to fire first. Do not build a pitch for a product that does not exist.
• Verdict: Ignore-until-trigger. Trigger: an AI-native product launch with sign-off consequences, plus a Staff+ requisition.

2. Betterment. Automated investing and retirement accounts for retail investors and advisors.

• Stage / Size / Capital: Late-stage private.
• Offices: New York.
• Value chain: Downstream consumer finance. Design owns the trust relationship with someone's savings.
• Best at: Making automated portfolio management legible to people who do not want to learn finance.
• Signals: No qualifying senior product design role on the careers site (verified Aug 1).
• Tier: Healthcare/Regulated (high-consequence evaluation).
• Rubric: Company 6/12 — AI 1, Stage/equity 2, Design influence 2, Trajectory 1. Role 8/15 — Comp 2, Scope 2, Craft 2, AI exposure 1, Portfolio 1.
• Portfolio match: An AI advisor changing allocations autonomously would be a delegation surface with retirement money at stake. Strong conceptually, weak against current reality, and New York compounds it.
• Verdict: Ignore-until-trigger. Trigger: an autonomous advisory product launch plus US-remote or Bay Area design hiring.

3. Justworks. Professional employer organization handling payroll, benefits, and compliance for small businesses.

• Stage / Size / Capital: Late-stage private.
• Offices: New York.
• Value chain: Downstream HR and compliance service. Design influence moderate; the product is largely process.
• Best at: Compliance coverage small businesses cannot maintain themselves.
• Signals: No qualifying senior product design role on the Greenhouse board (verified Aug 1).
• Tier: Healthcare/Regulated (high-consequence evaluation).
• Rubric: Company 4/12 — AI 1, Stage/equity 1, Design influence 1, Trajectory 1. Role 7/15 — Comp 2, Scope 1, Craft 2, AI exposure 1, Portfolio 1.
• Portfolio match: No defensible match at current product state. It is here for completeness of the monitoring grid, not because I expect it to convert.
• Verdict: Ignore-until-trigger. Deprioritize. Trigger: an AI compliance product with automated determination plus Bay Area design hiring. Both unlikely.

4. Flexport. Freight forwarding and supply chain software with visibility across ocean, air, and customs.

• Stage / Size / Capital: Late-stage private.
• Offices: San Francisco.
• Value chain: Vertically integrated logistics, service plus software. Design owns exception handling, where the value sits.
• Best at: Making a container's status legible across a dozen parties who share no systems.
• Signals: No qualifying senior product design role on the careers site (verified Aug 1).
• Tier: Enterprise platform.
• Rubric: Company 7/12 — AI 2, Stage/equity 2, Design influence 2, Trajectory 1. Role 9/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 1.
• Portfolio match: Supply-chain exception handling is multi-party governance under time pressure. A shipment held at customs, six parties with partial information, one person who has to decide. Thermo Fisher mySupply is the closest published proof, and the pharma-supply context makes it more than a loose analogy. Differentiator: you have coordinated six partners onto a single view in a regulated supply chain, which is Flexport's structural problem exactly.
• Verdict: Ignore-until-trigger. Good portfolio fit waiting on company trajectory. Trigger: an agentic exception-resolution product launch plus a Staff+ requisition.

5. Deel. Global payroll, contractor management, and employer-of-record services.

• Stage / Size / Capital: Late-stage private, fast-growing.
• Offices: Distributed, San Francisco presence.
• Value chain: Downstream global employment infrastructure. Design owns compliance-heavy onboarding surfaces.
• Best at: Compliant hiring in jurisdictions most companies cannot navigate.
• Signals: Unverified — direct board state not certified this pass.
• Tier: Healthcare/Regulated (high-consequence evaluation).
• Rubric: Company 7/12 — AI 1, Stage/equity 3, Design influence 1, Trajectory 2. Role 8/15 — Comp 2, Scope 2, Craft 2, AI exposure 1, Portfolio 1.
• Portfolio match: Cross-jurisdiction compliance rendered as an onboarding flow. Red Cross's consolidation is the mechanism, but AI exposure of 1 means the work would not serve your portfolio goals even if you won it.
• Verdict: Ignore-until-trigger. Trigger: AI moving into compliance determination, not just document processing.

6. Rippling. Unified HR, IT, and finance platform built on one employee data model.

• Stage / Size / Capital: Late-stage private, very large.
• Offices: San Francisco.
• Value chain: Vertically integrated workforce platform. Design influence constrained by an aggressive product-and-engineering culture.
• Best at: One data model driving provisioning across dozens of downstream systems.
• Signals: Unverified — direct board state not certified this pass.
• Tier: Enterprise platform.
• Rubric: Company 7/12 — AI 1, Stage/equity 3, Design influence 1, Trajectory 2. Role 10/15 — Comp 3, Scope 2, Craft 2, AI exposure 1, Portfolio 2.
• Portfolio match: Agents that provision and deprovision access autonomously would be an identity-delegation surface with real blast radius. Until then the influence ceiling is the binding problem, and you should raise it directly in any conversation here, because the public evidence suggests design executes rather than shapes.
• Verdict: Ignore-until-trigger. Trigger: an autonomous workforce-agent launch plus a Director+ requisition with a stated reporting line to product leadership.

7. Asana. Work management platform with AI teammates that take on assigned tasks.

• Stage / Size / Capital: Public.
• Offices: San Francisco.
• Value chain: Downstream productivity application. Design influence is high and long-established.
• Best at: Modeling work as structured objects rather than lists.
• Signals: No qualifying senior product design role in the reserve board pass (verified Aug 1).
• Tier: Enterprise platform.
• Rubric: Company 7/12 — AI 2, Stage/equity 1, Design influence 3, Trajectory 1. Role 10/15 — Comp 2, Scope 2, Craft 3, AI exposure 2, Portfolio 1.
• Portfolio match: AI teammates raise an accountability question no work tool has answered. When an agent misses a deadline, who does the system say is responsible? Your five-handoffs framework treats accountability transfer as a named handoff. Differentiator: you have vocabulary for a problem Asana's product team is currently describing in metaphors.
• Verdict: Ignore-until-trigger. Trigger: a Staff+ or Director requisition owning AI teammates specifically.

8. Lattice. Performance management, engagement surveys, and HR analytics.

• Stage / Size / Capital: Late-stage private.
• Offices: San Francisco.
• Value chain: Downstream HR application. Design owns a surface where managers make judgments about people.
• Best at: Making performance review cycles something managers complete rather than abandon.
• Signals: No qualifying senior product design role in the reserve board pass (verified Aug 1).
• Tier: Enterprise platform.
• Rubric: Company 7/12 — AI 2, Stage/equity 2, Design influence 2, Trajectory 1. Role 8/15 — Comp 2, Scope 2, Craft 2, AI exposure 1, Portfolio 1.
• Portfolio match: AI-assisted performance feedback is a consequential-judgment surface about employees, with legal exposure and heavy ethical weight. Structurally your kind of problem, commercially a small one. Company score is the constraint.
• Verdict: Ignore-until-trigger. Trigger: an AI evaluation product with employment consequences plus a Director+ requisition.

9. Plaid. Financial data connectivity between consumer bank accounts and fintech applications.

• Stage / Size / Capital: Late-stage private.
• Offices: San Francisco.
• Value chain: Midstream financial data infrastructure. Design owns the consumer consent surface, which carries more consequence than its size suggests.
• Best at: The bank-linking flow. Millions of consumers have used it without ever learning the company's name.
• Signals: Unverified — direct board state not certified this pass.
• Tier: Healthcare/Regulated (high-consequence evaluation).
• Rubric: Company 6/12 — AI 1, Stage/equity 2, Design influence 2, Trajectory 1. Role 9/15 — Comp 2, Scope 2, Craft 3, AI exposure 1, Portfolio 1.
• Portfolio match: Consent for AI agents to reach financial accounts is the coming version of Plaid's core surface: a consumer granting standing permission to something that will act, not merely read. Differentiator: you hold a public position on staged permission, which is precisely what that surface will need.
• Verdict: Ignore-until-trigger. Trigger: an agent-consent product launch plus a Staff+ requisition.

10. Coda. Document-database hybrid for team workflows, now part of Grammarly's productivity suite.

• Stage / Size / Capital: Acquired; part of a combined productivity company.
• Offices: San Francisco.
• Value chain: Downstream productivity application inside a larger suite. Design authority depends on suite integration decisions.
• Best at: Blurring the line between a doc and an app without teaching users a new model.
• Signals: Unverified — direct board state not certified this pass.
• Tier: Growth-stage platform.
• Rubric: Company 6/12 — AI 2, Stage/equity 1, Design influence 2, Trajectory 1. Role 9/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 1.
• Portfolio match: Suite consolidation is the live problem. Merge distinct products into one coherent surface without losing the users who chose one of them. Equinox+ consolidated five brands into a single MVP holding 4.8 stars, which is specific evidence that consolidation need not cost satisfaction. Differentiator: you can show a merged product that was better received than its parts.
• Verdict: Ignore-until-trigger. Trigger: a design leadership requisition naming the combined product surface.

7. Regulated and clinical, waiting for a decision-surface mandate

Corrected scoring makes this the most valuable cluster on the list, not the slowest one. Three of the eight highest role scores anywhere sit here: Viz.ai at 14, the single highest on the board and one company-rubric point off Act, plus Regard and Commure at 13. Veracyte at 12 is right behind. Mechanism reliability: pattern-inferred, long lead time, usually tied to a product clearing regulatory review or a health-system contract forcing a workflow change. Three of these run on the same Red Cross non-expert mechanism, so each gets a different decision from that case: uncertainty disclosure at Omada, matching legibility at Spring, crisis escalation at Lyra in Cluster 10. If more than one goes live, prioritize on company score.

1. Viz.ai. AI-powered care coordination. Detects stroke and cardiac events in imaging and routes care teams.

• Stage / Size / Capital: Late-stage private.
• Offices: San Francisco.
• Value chain: Downstream clinical decision support. Maximum consequence: the product routes a stroke patient to a specialist.
• Best at: Closing the gap between a radiologist reading a scan and a neurologist being paged.
• Signals: Unverified — direct board state not certified this pass.
• Tier: Healthcare/Regulated.
• Rubric: Company 9/12 — AI 3, Stage/equity 3, Design influence 2, Trajectory 1. Role 14/15 — Comp 2, Scope 3, Craft 3, AI exposure 3, Portfolio 3.
• Portfolio match: The care-team notification surface, a cardiologist getting an AI-flagged case on a phone at 3 a.m., is a time-critical alert with a single required action. Retail Velocity's field-alert triage is the mechanism, and the transferable decision is the escalation threshold: what earned an interruption versus what waited for the next review. Differentiator: you have already set an interrupt threshold in a live alerting product and had to defend where you drew the line. Bring the reasoning, not a metric, because the Agentic Labs work is not published with one.
• Verdict: Ignore-until-trigger. Highest role score on the list. Verify the board this week. Trigger: any Staff+ or Director product design requisition.

2. Regard. AI that reviews patient charts and surfaces potential diagnoses for physician confirmation.

• Stage / Size / Capital: Growth-stage private.
• Offices: Los Angeles; distributed.
• Value chain: Downstream clinical application inside the electronic health record. A physician confirms or rejects an AI-suggested diagnosis, which is as much decision weight as design gets anywhere.
• Best at: Living inside the existing chart workflow instead of asking clinicians to leave it.
• Signals: No qualifying senior product design role in the reserve board pass (verified Aug 1).
• Tier: Healthcare/Regulated.
• Rubric: Company 9/12 — AI 3, Stage/equity 3, Design influence 2, Trajectory 1. Role 13/15 — Comp 1, Scope 3, Craft 3, AI exposure 3, Portfolio 3.
• Portfolio match: A physician accepting an AI-proposed diagnosis into a legal medical record is the highest-stakes sign-off surface on this list. Your trust ladder separates Verify from Delegate precisely because some decisions must never reach Delegate, and this is the clearest instance of that distinction anywhere on the board. Differentiator: you have a published framework naming the exact failure mode this product must avoid, and you can cite it in a first conversation without sounding rehearsed.
• Verdict: Ignore-until-trigger. Note comp score 1. This is a function-founding role at growth stage, and cash will be the negotiation. Trigger: any product design requisition at any level.

3. Commure. AI platform for clinical operations covering ambient documentation, prior authorization, and care coordination.

• Stage / Size / Capital: Late-stage private.
• Offices: Mountain View.
• Value chain: Midstream clinical AI infrastructure with downstream workflow applications. Design influence rises in the workflow layer.
• Best at: Ambient documentation, capturing a clinical encounter without the clinician typing.
• Signals: No qualifying senior product design role in the reserve board pass (verified Aug 1).
• Tier: Healthcare/Regulated.
• Rubric: Company 9/12 — AI 3, Stage/equity 3, Design influence 2, Trajectory 1. Role 13/15 — Comp 2, Scope 3, Craft 3, AI exposure 3, Portfolio 2.
• Portfolio match: Prior authorization is a multi-party approval chain where an AI recommendation has to survive a payer's review. Same accountability-transfer problem as Carrier IQ, with a health outcome attached. Differentiator: you have designed approval chains where the machine's recommendation had to be legible to a reviewer who did not build the model.
• Verdict: Ignore-until-trigger. Trigger: a Staff+ or Director requisition naming workflow applications rather than infrastructure.

4. Veracyte. Genomic diagnostic tests with machine-learning classifiers for cancer diagnosis.

• Stage / Size / Capital: Public.
• Offices: South San Francisco.
• Value chain: Vertically integrated diagnostics from assay through report. Design owns the report clinicians act on.
• Best at: Reducing unnecessary surgeries by resolving indeterminate biopsy results.
• Signals: No qualifying senior product design role in the reserve board pass (verified Aug 1).
• Tier: Healthcare/Regulated.
• Rubric: Company 7/12 — AI 3, Stage/equity 1, Design influence 2, Trajectory 1. Role 12/15 — Comp 2, Scope 2, Craft 2, AI exposure 3, Portfolio 3.
• Portfolio match: The diagnostic report is the design surface. A classifier score a physician uses to recommend surgery or not, presented once, with no interaction. Carrier IQ is the closest analogue, and name the difference honestly in any conversation: there is no override loop here, only presentation. Differentiator: a publishable case study if it converts, which is why portfolio value scores 3 against a modest company score.
• Verdict: Ignore-until-trigger. Small design org, so the first senior requisition is the entire opportunity. Trigger: any Staff+ or Director product design requisition.

5. Spring Health. Mental health benefits platform connecting employees to therapists, coaches, and digital tools.

• Stage / Size / Capital: Late-stage private.
• Offices: New York.
• Value chain: Downstream mental health marketplace. Design owns member-facing care navigation.
• Best at: Matching a member to the right level of care, not just to an available therapist.
• Signals: No qualifying senior product design role in the reserve board pass (verified Aug 1).
• Tier: Healthcare/Regulated.
• Rubric: Company 8/12 — AI 2, Stage/equity 3, Design influence 2, Trajectory 1. Role 11/15 — Comp 2, Scope 2, Craft 3, AI exposure 2, Portfolio 2.
• Portfolio match: AI-driven care matching is a recommendation surface where the member cannot evaluate the recommendation. The Red Cross decision that transfers here is matching legibility: showing the basis for an assignment so a person under stress accepts it instead of restarting the process. Differentiator: "designed for vulnerable users under stress" is a narrower and far more credible claim than "consumer health design."
• Verdict: Ignore-until-trigger. Note New York geography. Trigger: a Staff+ or Director requisition.

6. Hinge Health. Digital physical therapy and musculoskeletal care with wearable sensors and coaching.

• Stage / Size / Capital: Public.
• Offices: San Francisco.
• Value chain: Downstream care delivery. Design owns the patient-facing exercise and coaching surface.
• Best at: Keeping patients in a physical therapy program they would otherwise abandon after two sessions.
• Signals: Unverified — direct board state not certified this pass.
• Tier: Healthcare/Regulated.
• Rubric: Company 7/12 — AI 2, Stage/equity 2, Design influence 2, Trajectory 1. Role 11/15 — Comp 2, Scope 2, Craft 3, AI exposure 2, Portfolio 2.
• Portfolio match: AI-personalized exercise progression, where the system adjusts a plan from sensor data with no therapist in the loop, is a delegation surface with a pain outcome. Equinox+ is your consumer-health proof of building a program people finish. Differentiator: you shipped a multi-brand health product users rated 4.8 stars, which is the retention proof this company needs.
• Verdict: Ignore-until-trigger. Trigger: verify the board, then watch for a Staff+ or Director requisition owning the patient experience.

7. Komodo Health. Healthcare data platform mapping patient journeys across claims and clinical data.

• Stage / Size / Capital: Late-stage private.
• Offices: San Francisco.
• Value chain: Midstream healthcare data. Design owns the analysis surface for life-sciences and payer customers.
• Best at: Longitudinal patient-journey data at national scale.
• Signals: No qualifying senior product design role on the Greenhouse board (verified Aug 1).
• Tier: Healthcare/Regulated.
• Rubric: Company 7/12 — AI 2, Stage/equity 2, Design influence 2, Trajectory 1. Role 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 2.
• Portfolio match: Cohort construction is the design problem. A researcher building a patient population needs to see how each inclusion rule changed the result. Iterative query legibility, closest to your Alibaba search work in mechanism and nothing like it in domain. Differentiator: you have made a ranking system's logic visible to non-technical users at scale.
• Verdict: Ignore-until-trigger. Trigger: a Staff+ or Director requisition naming an AI-driven analysis product.

8. PointClickCare. Electronic health records and care coordination for post-acute and long-term care.

• Stage / Size / Capital: Late-stage private.
• Offices: Mississauga, Ontario; US presence.
• Value chain: Downstream clinical records for a segment most EHR vendors underserve. Design owns the care-team workflow.
• Best at: Serving skilled nursing facilities and home health agencies that Cerner and Epic ignore.
• Signals: Unverified — direct board state not certified this pass.
• Tier: Healthcare/Regulated.
• Rubric: Company 7/12 — AI 2, Stage/equity 2, Design influence 2, Trajectory 1. Role 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 2.
• Portfolio match: Care-transition documentation, a nurse discharging a patient to home health and needing the receiving team to hold the right information, is a multi-party handoff with a safety consequence. No differentiation here. Your five-handoffs framework applies, and so does any senior healthcare designer's discharge-flow experience.
• Verdict: Ignore-until-trigger. Low priority: Canadian HQ, unverified board, no positioning edge. Trigger: verify the board, then a US-eligible Staff+ requisition.

9. Flatiron Health. Oncology electronic health records and real-world evidence for cancer research.

• Stage / Size / Capital: Roche subsidiary.
• Offices: New York, San Francisco.
• Value chain: Vertically integrated clinical software plus research data. Corporate ownership caps the equity thesis hard.
• Best at: Turning messy oncology chart data into research-grade evidence.
• Signals: No qualifying senior product design role on the careers page (verified Aug 1).
• Tier: Healthcare/Regulated.
• Rubric: Company 6/12 — AI 2, Stage/equity 1, Design influence 2, Trajectory 1. Role 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 2.
• Portfolio match: Oncology abstraction review: a trained abstractor confirms or corrects what a model extracted from a chart, and an error propagates into published research. Carrier IQ is the mechanism, and the distinction worth naming is that both surfaces correct machine extraction rather than approve machine recommendation. Differentiator: you designed the correction loop, not just the approval button.
• Verdict: Ignore-until-trigger. Roche ownership makes equity here compensation, not upside. Trigger: a Staff+ or Director design requisition.

10. Omada Health. Virtual care for chronic conditions including diabetes, hypertension, and musculoskeletal.

• Stage / Size / Capital: Public.
• Offices: San Francisco.
• Value chain: Downstream care delivery. Design owns the patient behavior-change surface.
• Best at: Sustained engagement in chronic-condition programs, where most digital health drops off.
• Signals: No qualifying senior product design role on the Greenhouse board (verified Aug 1).
• Tier: Healthcare/Regulated.
• Rubric: Company 6/12 — AI 2, Stage/equity 1, Design influence 2, Trajectory 1. Role 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 2.
• Portfolio match: AI coaching for a chronic-condition patient turns on one decision: when the system says it does not know enough and tells the patient to call a doctor. That is the uncertainty-disclosure decision from Red Cross, where the job was admitting the system's limits without collapsing the user's confidence in it. Differentiator: you have designed the "I can't help with this" state for a high-stakes system, which most consumer health products still treat as an error page.
• Verdict: Ignore-until-trigger. Trigger: a Director+ requisition plus evidence AI has moved from coaching support to clinical recommendation.

8. Autonomy in the physical world

Ranked by corrected score, which puts Anduril and Zipline above Waymo. Every company here builds, or built, a system that acts physically with no human in the loop, and every one has a supervision surface that is either unbuilt or built badly. Mechanism reliability: pattern-inferred. The trigger is nearly always a regulatory milestone: an operational expansion, a safety certification, a government contract requiring a documented human-oversight interface. Three of these run on Retail Velocity, so I have split it into three separable decisions: grouping logic at Zipline, escalation threshold at Waymo, live re-prioritization at Verkada.

1. Anduril. Defense technology: autonomous systems, sensor fusion, and command software for the US military.

• Stage / Size / Capital: Late-stage private, very large.
• Offices: Costa Mesa, Washington DC, distributed.
• Value chain: Vertically integrated defense platform. Design owns the command-and-control surface where human judgment meets autonomous action.
• Best at: Shipping defense hardware and software at a commercial pace instead of a procurement pace.
• Signals: No qualifying senior product design role on the open roles page (verified Aug 1).
• Tier: Healthcare/Regulated (high-consequence evaluation).
• Rubric: Company 10/12 — AI 3, Stage/equity 3, Design influence 2, Trajectory 2. Role 13/15 — Comp 2, Scope 3, Craft 3, AI exposure 3, Portfolio 2.
• Portfolio match: The command surface is a human authorizing an autonomous system to act in a contested environment. The highest-stakes version of the Delegate rung. Differentiator: your published framework names the conditions under which delegation is appropriate, and this design team would take that conversation seriously. Say it plainly: defense work carries ethical weight, and that is a decision to make before you pursue it, not during an interview.
• Verdict: Ignore-until-trigger. Trigger: a Staff+ or Director requisition naming command software or human-machine teaming.

2. Zipline. Autonomous drone delivery for medical supplies, retail, and food across four continents.

• Stage / Size / Capital: Late-stage private.
• Offices: South San Francisco.
• Value chain: Vertically integrated across drone, logistics software, and distribution network. Design owns the operations console and customer notification.
• Best at: Delivering blood products to rural hospitals in under thirty minutes. The medical use case proves everything else.
• Signals: Unverified — direct board state not certified this pass.
• Tier: Healthcare/Regulated (high-consequence evaluation).
• Rubric: Company 10/12 — AI 3, Stage/equity 3, Design influence 2, Trajectory 2. Role 13/15 — Comp 2, Scope 3, Craft 3, AI exposure 3, Portfolio 2.
• Portfolio match: An operations console managing hundreds of simultaneous flights is a volume problem before it is a judgment problem. The transferable Retail Velocity decision is grouping logic: batching alerts by underlying cause instead of by timestamp, which turned a queue no human could read into a queue one operator could work. Differentiator: you made that grouping call in a shipped alerting product, which is a different conversation from having opinions about queue design.
• Verdict: Ignore-until-trigger. Trigger: verify the board, then watch for a Staff+ or Director requisition naming operations or logistics software.

3. Applied Intuition. Simulation and tooling for autonomous vehicle development, used by most major AV programs.

• Stage / Size / Capital: Late-stage private, very large.
• Offices: Sunnyvale.
• Value chain: Midstream AV development infrastructure. Design owns the simulation console and scenario-authoring surface.
• Best at: Making AV simulation fast enough to live inside a daily development loop.
• Signals: Unverified — direct board state not certified this pass.
• Tier: Enterprise platform.
• Rubric: Company 10/12 — AI 3, Stage/equity 3, Design influence 2, Trajectory 2. Role 12/15 — Comp 3, Scope 2, Craft 3, AI exposure 3, Portfolio 1.
• Portfolio match: Scenario authoring is a domain-expert tool problem. An autonomy engineer builds a test case for a rare edge condition. Brand Pulse's attribution surface is the closest published analogue for expert-tool design where output has to be auditable. Differentiator: you have built for experts who will not accept a black box, which is the harder version of trust design.
• Verdict: Ignore-until-trigger. Company 10 and high-value if the board is clean. Verify before building a pitch. Trigger: a Staff+ or Director requisition.

4. Scale AI. Data labeling, evaluation, and AI application development for enterprise and government.

• Stage / Size / Capital: Late-stage private, very large.
• Offices: San Francisco.
• Value chain: Midstream AI infrastructure covering data quality and model evaluation. Design influence rises in the enterprise application layer.
• Best at: Human-in-the-loop data quality at a scale that makes model training tractable.
• Signals: No recent signal — Scale AI named Francis deSouza CEO on July 30, 2026. A CEO appointment falls outside the tracked trigger categories and is not a design-window signal. Board state not certified.
• Tier: Enterprise platform.
• Rubric: Company 10/12 — AI 3, Stage/equity 3, Design influence 2, Trajectory 2. Role 11/15 — Comp 3, Scope 2, Craft 2, AI exposure 3, Portfolio 1.
• Portfolio match: The evaluation surface, a human reviewer rating model outputs against a rubric at volume, is a design problem Scale AI has not solved publicly. Brand Pulse is the quality-judgment analogue. Differentiator: you have designed human-review workflows at volume. That sits at the center of this company's operations, not off to the side.
• Verdict: Ignore-until-trigger. Trigger: a Staff+ or Director requisition. The leadership change is context, not a trigger, and I am being explicit because it will be tempting to treat it as one.

5. Waymo. Fully autonomous ride-hailing operating commercially in San Francisco, Phoenix, and Los Angeles.

• Stage / Size / Capital: Alphabet subsidiary; very large.
• Offices: Mountain View, San Francisco.
• Value chain: Vertically integrated from sensor stack through rider app. Corporate ownership caps equity; design influence is real in the rider and operator surfaces.
• Best at: Commercial autonomous operation at scale. Nobody else is close.
• Signals: No qualifying senior product design role on the careers site (verified Aug 1). Figma's Config 2026 speaker material identifies Ryan Powell as Head of Design.
• Tier: Healthcare/Regulated (high-consequence evaluation).
• Rubric: Company 9/12 — AI 3, Stage/equity 1, Design influence 2, Trajectory 3. Role 12/15 — Comp 2, Scope 2, Craft 3, AI exposure 3, Portfolio 2.
• Portfolio match: Remote assistance, an operator intervening in a vehicle situation from a console miles away, turns entirely on the escalation threshold. That is the Retail Velocity decision that transfers here. You have had to decide what earns an interruption and what the system absorbs silently. Differentiator: you can walk a panel through that trade-off from a shipped product rather than from theory.
• Verdict: Ignore-until-trigger. Trigger: a Staff+ or Director requisition under Powell, or a public change to his seat.

6. Shield AI. Autonomous AI pilot for military aircraft and drones, operating without GPS or communications.

• Stage / Size / Capital: Late-stage private.
• Offices: San Diego, Washington DC.
• Value chain: Vertically integrated defense AI. Design owns the mission-planning and debrief surfaces.
• Best at: Flying a fighter aircraft autonomously in a denied environment. The technical bar is genuinely extraordinary.
• Signals: Unverified — direct board state not certified this pass.
• Tier: Healthcare/Regulated (high-consequence evaluation).
• Rubric: Company 9/12 — AI 3, Stage/equity 3, Design influence 1, Trajectory 2. Role 11/15 — Comp 2, Scope 2, Craft 2, AI exposure 3, Portfolio 2.
• Portfolio match: Mission debrief is decision-sequence reconstruction. A pilot reviews what the AI did and why, after the fact, with no ability to intervene. Carrier IQ's audit trail is the mechanism, and the distinction worth naming is that this surface is retrospective rather than prospective. Differentiator: you have designed accountability artifacts for autonomous decisions, not dashboards around them.
• Verdict: Ignore-until-trigger. Same ethical-weight caveat as Anduril. Trigger: a Staff+ or Director requisition naming mission planning or human-machine interface.

7. Skydio. Autonomous drones for inspection, public safety, and defense.

• Stage / Size / Capital: Late-stage private.
• Offices: San Mateo.
• Value chain: Vertically integrated across drone hardware, autonomy software, and fleet management. Design owns the operator console.
• Best at: Obstacle avoidance. The autonomy stack is genuinely best-in-class for close-proximity inspection.
• Signals: Unverified — direct board state not certified this pass.
• Tier: Healthcare/Regulated (high-consequence evaluation).
• Rubric: Company 8/12 — AI 3, Stage/equity 2, Design influence 2, Trajectory 1. Role 11/15 — Comp 2, Scope 2, Craft 2, AI exposure 3, Portfolio 2.
• Portfolio match: The inspection-review surface, a bridge engineer reviewing AI-flagged anomalies from a drone pass, is a correction-loop problem. The human's disagreement has to be recorded and fed back. Carrier IQ's correction loop is the mechanism here, not the approval gate. Differentiator: you have designed the surface where a human's disagreement is captured as structured input instead of discarded.
• Verdict: Ignore-until-trigger. Trigger: verify the board, then a Staff+ or Director requisition naming fleet management or inspection software.

8. Verkada. Cloud-managed physical security: cameras, access control, and environmental sensors.

• Stage / Size / Capital: Late-stage private.
• Offices: San Mateo.
• Value chain: Vertically integrated hardware plus software for physical security. Design owns the security operations console.
• Best at: Making enterprise physical security manageable without a dedicated security team.
• Signals: Unverified — direct board state not certified this pass.
• Tier: Enterprise platform.
• Rubric: Company 8/12 — AI 2, Stage/equity 3, Design influence 2, Trajectory 1. Role 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio 2.
• Portfolio match: AI-flagged security events, a camera detecting unusual behavior and surfacing it to a guard, is a live re-prioritization problem. The queue changes while the operator is working it. That is the third separable Retail Velocity decision. Not grouping, not threshold, but the moment the ranked list updates mid-review. Differentiator: you have designed for a queue that moves, which is harder than designing for one that waits.
• Verdict: Ignore-until-trigger. Trigger: verify the board, then a Staff+ or Director requisition.

9. Cruise. Formerly an autonomous ride-hailing operator; the autonomy work now sits inside GM's driver-assistance program.

• Stage / Size / Capital: Wholly owned by GM.
• Offices: San Francisco; parent HQ Detroit.
• Value chain: Autonomy stack absorbed into an automaker's product line. Value concentrates in GM's vehicle programs, so an OEM release cycle sets design scope.
• Best at: Historically, dense-urban driverless operation. That mandate no longer exists.
• Signals: No recent signal — GM acquired full ownership and redirected the program away from robotaxi development on February 4, 2025. Ownership and operating thesis both changed; board state not certified this pass.
• Tier: Healthcare/Regulated (high-consequence evaluation).
• Rubric: Company 4/12 — AI 2, Stage/equity 1, Design influence 1, Trajectory 1. Role 7/15 — Comp 2, Scope 1, Craft 2, AI exposure 1, Portfolio 1.
• Portfolio match: None worth constructing. Driver-assistance handoff design inside an automaker is real work, but it sits under a vehicle program rather than a product org, and nothing published of yours speaks to that reporting reality.
• Verdict: Ignore-until-trigger. The original thesis does not come back. Trigger: a GM requisition naming autonomous or driver-assistance interface ownership at Staff+ or Director with disclosed scope. Low expectation, and I would rather you knew that than kept the tab open.

9. The IPO window

Companies that have filed, priced, or are credibly expected to price, plus recent pricings whose lock-ups have already run. The equity thesis is real and time-bounded: the stretch between S-1 and lock-up expiration is the only period where a senior design hire captures meaningful pre-liquidity upside. Mechanism reliability: high on the companies that have filed; speculative on the expected-to-price entries. Thin cluster by design this cycle. Chime routed to the active queue on a live Director, Product Design mandate, and Klarna and Navan are scored in Cluster 1, where the capital event rather than the window is the governing fact. The trigger here never changes: a Staff+ or Director requisition disclosing RSU grant size and vesting schedule. Without that disclosure the equity story is a rumor.

1. Cerebras. Custom wafer-scale AI chips and a cloud inference service.

• Stage / Size / Capital: Private. Withdrew its S-1 in October 2025 after a large private round; no active filing as of Aug 1.
• Offices: Sunnyvale.
• Value chain: Upstream inference infrastructure. Design influence structurally low.
• Best at: Raw compute density. The wafer-scale chip is a genuine engineering achievement.
• Signals: Unverified — direct board state not certified this pass.
• Tier: AI-native.
• Rubric: Company 7/12 — AI 3, Stage/equity 2, Design influence 1, Trajectory 1. Role 8/15 — Comp 2, Scope 1, Craft 2, AI exposure 2, Portfolio 1.
• Portfolio match: Developer console for a novel compute substrate. Thin, and the infrastructure layer caps influence regardless.
• Verdict: Ignore-until-trigger. Trigger: a refiled S-1 plus a Head/Director design requisition with disclosed equity terms.

2. Coreweave. GPU cloud infrastructure for AI workloads, now public.

• Stage / Size / Capital: Public; IPO completed March 28, 2025.
• Offices: Roseland, NJ; distributed.
• Value chain: Upstream compute infrastructure. Design influence at this layer is reliably low.
• Best at: GPU availability at a moment when hyperscalers are rationing.
• Signals: Unverified — direct board state not certified this pass.
• Tier: AI-native.
• Rubric: Company 6/12 — AI 2, Stage/equity 2, Design influence 1, Trajectory 1. Role 7/15 — Comp 2, Scope 1, Craft 2, AI exposure 1, Portfolio 1.
• Portfolio match: None specific. Infrastructure layer, developer audience, no design surface worth naming.
• Verdict: Ignore-until-trigger. Trigger: a design requisition naming an application surface rather than a console.

3. Klarna. Already listed in Cluster 1, entry 9. Not duplicated here.

4. Navan. Already listed in Cluster 1, entry 8. Not duplicated here.

10. Reason to look, no certified board state

Ten companies where I hold a signal (a funding event, a product milestone, a leadership change, a market position) that argues something might be live, and where I did not certify the board this pass. Cheapest information gain on the list. Two minutes each. No board links here, by rule: I do not publish a URL I did not load. The addresses below are where to start looking, not confirmed postings. If a path has moved, search the company name plus "careers." Ranked 1 through 10 by expected value, meaning company score times my estimate of the probability a qualifying role exists.

1. Mercury. Banking for startups: checking, savings, treasury, and venture debt on one platform.

• Signal: Company 10, craft-forward culture, no certified board state. Highest expected value of any unverified entry.
• Verify at: mercury.com/company/careers
• Trigger: any Staff+ or Director product design requisition.

2. Hex. Collaborative data notebooks for analysts and data scientists, with AI-assisted query generation.

• Signal: Late-stage private, AI-native data tool, no certified board state.
• Verify at: hex.tech/company/careers
• Trigger: a Staff+ or Head of Design requisition.

3. Retool. Internal tool builder for developers, expanding into AI-assisted app generation.

• Signal: Late-stage private, developer platform moving into the AI application layer.
• Verify at: retool.com/careers
• Trigger: a Staff+ or Director requisition naming the AI application surface.

4. Weights and Biases. ML experiment tracking and model management, now part of CoreWeave following the 2025 acquisition.

• Signal: The independent equity thesis is gone and any design mandate routes through a public parent. Monitor CoreWeave's board, not wandb's.
• Verify at: coreweave.com/careers
• Trigger: a CoreWeave requisition naming W&B experiment or evaluation surfaces at Staff+ or Director.

5. Temporal. Workflow orchestration platform for durable execution of long-running processes.

• Signal: Late-stage private, developer infrastructure with an emerging application layer.
• Verify at: temporal.io/careers
• Trigger: a Staff+ or Head of Design requisition naming an application surface.

6. Lyra Health. Mental health benefits platform for enterprise employees, with clinical care and coaching.

• Signal: Late-stage private, healthcare/regulated, no certified board state. The crisis-escalation Red Cross decision belongs here if a trigger fires.
• Verify at: lyrahealth.com/about/careers
• Trigger: a Staff+ or Director requisition.

7. Included Health. Integrated virtual care and navigation for health plan members.

• Signal: Late-stage private, healthcare/regulated. Honest note: I identified no differentiation this pass. Listed because the market position argues for a design mandate.
• Verify at: includedhealth.com/careers
• Trigger: a Staff+ or Director requisition plus a differentiation angle I can source.

8. Benchling. R-and-D data platform for life sciences, covering lab notebooks, sample management, and genomics.

• Signal: Late-stage private, regulated industry, complex multi-party data surface.
• Verify at: benchling.com/careers
• Trigger: a Staff+ or Director requisition naming a scientist-facing application surface.

9. Ironclad. AI-powered contract management for legal and procurement teams.

• Signal: Late-stage private, AI moving into a high-consequence sign-off surface.
• Verify at: ironcladapp.com/careers
• Trigger: a Staff+ or Director requisition naming the AI review or negotiation surface.

10. Watershed. Corporate carbon accounting and climate program management.

• Signal: Late-stage private, regulated reporting surface, no certified board state.
• Verify at: watershed.com/careers
• Trigger: a Staff+ or Director requisition naming the reporting or audit surface.