One posting worth your time this week. One more worth your time in the next two weeks. Five that aren't. Fieldguide (Staff Product Designer, 14 days old, remote USA, Series C) is Act. Ambience Healthcare (Staff Product Designer, 30 days old, SF, Series C) is Watch — the posting has crossed into its extended window, but healthcare domain specialization and a thin candidate pool justify continued pursuit. Contact Fieldguide first. Both companies build products where a licensed professional must decide whether an AI-generated output is trustworthy enough to sign their name to it. The other five companies in this scan have decayed past their windows or vanished from their ATS feeds.
HackerOne is the one that stings. Director of AI Product Design. Role rubric: 15/15, the highest in this entire set. The posting language read like someone reverse-engineered a job description from your published trust framework. It appears to have been pulled from HackerOne's active Ashby feed. Details and trigger conditions below.
The organizing filter across all seven: decision-rights density. How many moments per session does the product force a human to decide whether to trust, override, or delegate to an AI system? That density predicts portfolio value and scope expandability more reliably than brand recognition, headcount, or title. A company nobody mentions at Config whose product forces a human override decision every 90 seconds will produce a more publishable, more career-defining case study than a household name where AI appears on the roadmap slide and nowhere in the product. Fieldguide and Ambience sit at the top of that density curve. The five decayed signals are a reminder that the window matters as much as the fit.
Scanned 170+ companies across Ashby, Greenhouse, Lever, LinkedIn Jobs, and VC portfolio boards (a16z, Sequoia, Goldman Sachs, Kleiner Perkins, ICONIQ) before narrowing to these seven on rubric score and decision-rights density.
Window Status Changes This Cycle
- Downgraded: HackerOne (Act → Ignore-until-trigger; posting pulled from active Ashby feed, 49 days old). Abridge (Watch → Ignore-until-trigger; 325-day-old posting, zombie; previously flagged in Issue #4).
- Expiring: Ambience Healthcare (posting crosses 42-day Watch threshold in ~12 days; verify ATS status before next cycle).
Verification Infrastructure
The design problem that matters here: an AI generates an output. A licensed professional must decide whether that output is accurate enough to attach their name, their license, their liability to it. An audit finding. A clinical note. The designer's job is the judgment moment itself — making it faster, more confident, more defensible. Fieldguide is hot. Ambience is in its extended window with justified Watch status. Abridge, previously flagged in Issue #4 as the strongest healthcare AI trust signal worth monitoring, has been sitting at 325 days. That posting is a zombie.
-
Fieldguide. Software for assurance and audit practitioners across cybersecurity, privacy, and financial audits, with agentic AI embedded alongside human professionals throughout engagement execution.
- Stage / Size / Capital: Series C ($75M, Feb 2026, led by Goldman Sachs Alternatives). $125M total raised. $700M valuation. 51–200 employees.
- Offices: San Francisco HQ, remote-friendly. Role is remote, USA.
- Value chain: Midstream compliance tooling. Sits between the regulatory frameworks that define what must be audited and the enterprises being audited. Value concentrates in workflow automation and evidence management — the layer where AI can accelerate professional judgment without replacing professional accountability. Design scope is high because the product IS the workflow.
- Best at: Embedding AI into a domain where the human cannot legally delegate the final judgment. The auditor signs the opinion. The AI surfaces evidence, flags anomalies, drafts findings. The signature stays human. That constraint makes the design problem genuinely hard and structurally uncappable.
- Signals: Staff Product Designer posted Jul 10, 2026. 14 days old, hot. The posting language: "own complex ambiguous problem spaces end-to-end," "design AI-assisted workflows," "account for regulatory constraints," "influence senior PM and engineering leaders." Series C closed Feb 2026. Five months into the post-funding build phase. This posting is the design bet.
- Tier: Growth-stage platform with Healthcare/Regulated evaluation patterns. Tier Playbook: Read both Growth-stage platform and Healthcare/Regulated playbooks before outreach. The company's stage is growth (Series C, building out, design org forming), but the evaluation will center on regulatory-constraint thinking. Auditors bear personal liability for the outputs this product helps generate. Expect the interview to probe how you think about compliance constraints, high-consequence failure modes, and designing for users who are personally accountable for system outputs. The Growth-stage playbook prepares you for the function-building and player-coach questions. The Healthcare/Regulated playbook prepares you for the questions that will actually determine whether they hire you.
- Rubric scores: Company 11/12 — AI centrality 2 (AI-assisted, not AI-core), Stage/equity 3, Design influence 3 (end-to-end ownership, influences senior leaders), Trajectory 3. Role 14/15 — Total comp 2 (Staff level, no comp listed; Series C at $700M valuation suggests competitive), Scope 3 (end-to-end, regulatory domain, no structural ceiling), Craft depth 3 (hands-on IC, complex ambiguous problems), AI exposure 3 (AI-assisted workflows under regulatory constraints = trust and decision surfaces), Portfolio value 3.
- Portfolio match: Lead with Thermo Fisher mySupply. Fieldguide's hardest design problem is the auditor's sign-off moment: an AI surfaces a finding in a cybersecurity audit, and the practitioner must decide whether the evidence is sufficient to include it in a formal opinion that carries their professional liability. The specific match: mySupply's approval workflow where pharma supply chain managers signed off on system-generated recommendations knowing their name was on the compliance record. 100% adoption across 6 pharma partners. That proves you design for users who bear personal accountability for system outputs, at $20M+ margin, in production. You have designed for defensibility at scale in a regulated domain. That is the entire design problem at Fieldguide, and no other candidate walks in with that proof point at that scale.
- Verdict: Act. 14 days old, hot window. Company 11 ≥ 10, Role 14 ≥ 12. Remote USA. Contact this week. Lead with Thermo Fisher, bridge to TinyFish for AI-agent currency. Ask early about the relationship between this Staff IC role and the broader design function. "Influence senior PM and engineering leaders" is high-authority IC language. That phrase tells you more about the actual scope than the title does.
-
Ambience Healthcare. AI intelligence platform for healthcare delivering real-time coding-aware clinical documentation and workflow support across ambulatory, emergency, and inpatient settings, integrated with Epic, Oracle Cerner, and athenahealth.
- Stage / Size / Capital: Series C ($243M, Jul 2025, co-led by Oak HC/FT and a16z, with OpenAI Startup Fund and Kleiner Perkins). ~$313–345M total raised. 201–500 employees.
- Offices: San Francisco HQ. Role is San Francisco.
- Value chain: Downstream clinical application. Sits at the point of care where AI listens to the patient-clinician conversation and generates structured documentation in real time. Value concentrates in the accuracy and trustworthiness of the generated note, because that note enters the permanent medical record and drives billing, treatment decisions, and legal exposure.
- Best at: Real-time clinical AI at scale with multi-EHR integration. The integration depth across Epic, Cerner, and athenahealth is the moat. The product must work inside existing clinical workflows, not replace them. You are designing trust surfaces inside someone else's system. That constraint shapes everything.
- Signals: Staff Product Designer posted Jun 24, 2026. 30 days old, inside extended Watch window. The posting: "own end-to-end UX across Clinical AI products," partner with "clinical experts, customers, and clinical AI research teams." OpenAI Startup Fund participation in the Series C signals deep model-layer investment. High confidence this search is still active given the specialized domain and thin candidate pool for clinical AI design at this level.
- Tier: Healthcare/Regulated. Tier Playbook: Read Healthcare/Regulated playbook before outreach.
- Rubric scores: Company 11/12 — AI centrality 3, Stage/equity 3, Design influence 2 (Staff IC, end-to-end UX ownership but unclear strategic seat), Trajectory 3. Role 13/15 — Total comp 2 (Staff level, no comp listed), Scope 2 (clinical AI products, Staff level may have ceiling), Craft depth 3 (hands-on clinical AI), AI exposure 3 (real-time clinical documentation, trust, verification), Portfolio value 3.
- Portfolio match: Lead with "Trust Is the New Interface," specifically the Verify stage of the Watch→Verify→Delegate trust ladder. Ambience's hardest design problem is the clinician's review moment: an AI generates a clinical note from a live patient conversation, and the clinician must verify its accuracy before it enters the medical record. A hallucinated finding has direct patient-care consequences. Clinicians using Ambience are not watching passively and they are not delegating fully. They are verifying under time pressure with professional liability attached. Your published framework describes that exact calibration problem. Bridge to Agentic Labs Carrier IQ for production proof of auditable AI outputs mapped to source data. You walk in with a public, linkable artifact that describes the trust problem they are hiring someone to solve. No other candidate has that.
- Verdict: Watch. 30 days old. Posting is inside the 22–42 day extended window. Watch verdict justified by healthcare domain specialization, thin candidate pool for clinical AI design, and strong rubric scores (Company 11 ≥ 10, Role 13 ≥ 12). Contact within the next two weeks. Prioritize Fieldguide first — its posting is 16 days fresher and the role scope reads broader. Ask Ambience early about design's reporting line and whether "own end-to-end UX" means strategic influence or execution ownership. That distinction determines whether scope expandability is a 2 or a 3.
Ambience decay warning — Aug 5
This posting crosses the 42-day decay threshold around Aug 5. If you haven't made contact by then, verify the posting is still live in the Ashby feed before investing further.
- Abridge. Transforms patient-clinician conversations into structured clinical notes in real time, with Linked Evidence mapping AI-generated summaries to ground truth across 150+ enterprise health systems.
- Stage / Size / Capital: Series E ($300M, Jun 2025, led by a16z). 201–500 employees. 50M+ medical conversations projected in 2025, 55 specialties, 28 languages.
- Offices: San Francisco HQ, Pittsburgh, NYC. Builder hubs in SF and NYC. Role listed as hybrid, SF or NYC.
- Value chain: Downstream clinical application, same layer as Ambience but with a Linked Evidence differentiator that makes the auditability problem explicit in the product architecture.
- Best at: Linked Evidence as a product concept. Every AI-generated clinical statement traceable to the specific moment in the conversation that produced it. That is a design commitment, not a technical feature.
- Signals: Staff Product Designer published Sep 2, 2025. 325 days old. Still listed as active in the Ashby feed, but far past any reasonable decay window. Issue #4 flagged Abridge as worth watching for the first Director+ posting as clinical AI trust design matures there. That recommendation stands.
- Tier: Healthcare/Regulated.
- Rubric scores: Company 11/12 — AI centrality 3, Stage/equity 3, Design influence 2, Trajectory 3. Role 13/15 — Total comp 2, Scope 2, Craft depth 3, AI exposure 3, Portfolio value 3.
- Portfolio match: If a Director+ role appears, lead with Agentic Labs Carrier IQ. Abridge's Linked Evidence concept is structurally identical to the auditability challenge you solved in Carrier IQ: tracing AI-generated insurance quotes to source data so a professional user can verify before acting. Their product innovation is your portfolio precedent.
- Verdict: Ignore-until-trigger. 325 days old. The posting is either an evergreen pipeline role or a zombie. Trigger: new Staff or Director+ product design posting, or design leadership hire announcement. Abridge remains the single strongest healthcare AI trust match for your portfolio. The problem is timing, not fit.
Agent Governance
The highest decision-rights density in this scan sits where the design problem is who decides before an autonomous agent acts. An AI agent testing another AI system's defenses. An AI-driven financial workflow recommending a lending decision. Governance problems that need a designer. They produce the most differentiated portfolio work in the market right now. Neither posting in this cluster is live. HackerOne's disappearance is the more significant loss.
- HackerOne. Continuous Threat Exposure Management platform combining agentic AI with human security researchers to discover, validate, prioritize, and remediate security exposures across code, cloud, and AI systems.
- Stage / Size / Capital: Series E ($49M, Jan 2022). ~$160M total raised. 201–500 employees. Founded 2012.
- Offices: San Francisco HQ, with hubs in DC, Seattle, Austin, Boston, London, Groningen, Pune. Flexible-work. Role was listed as remote within ~50 mi of SF Bay Area, Boston, Seattle, Austin, DC, or London.
- Value chain: Midstream security platform. Sits between organizations that need to find vulnerabilities and the researchers (human and AI) who find them. Value concentrates in the orchestration layer: managing the interaction between autonomous AI agents and human security researchers, surfacing findings that drive remediation decisions. The trust architecture IS the product.
- Best at: The agent-to-agent trust surface. When an AI agent tests another AI system's defenses, the resulting findings must be interpretable, verifiable, and actionable by a human security team. Nobody else in the market is designing this interaction at this specificity.
- Signals: Director, AI Product Design posted ~Jun 5, 2026 per job mirror. 49 days old, and the role no longer appears in HackerOne's active Ashby feed. The posting described a Director reporting to the CPO, defining AI-native security workflows, agent-to-agent interactions, human-in-the-loop oversight, transparency, accountability, visualization, and interaction models. Comp: $243K–$300K + equity.
- Tier: AI-native.
- Rubric scores: Company 10/12 — AI centrality 3, Stage/equity 2 (Series E was Jan 2022; equity window may be narrowing without a recent round), Design influence 3 (Director reports to CPO), Trajectory 2 (stable, no recent funding signal). Role 15/15 — Total comp 3 ($243K–$300K + equity), Scope 3 (Director, AI-native security, no ceiling), Craft depth 3 (defining AI workflows, agent-to-agent), AI exposure 3 (agent-to-agent, human-in-the-loop, trust), Portfolio value 3.
- Portfolio match: The posting language — agent-to-agent interactions, human-in-the-loop oversight, transparency, accountability — reads like a job description reverse-engineered from "Trust Is the New Interface." The five-handoff framework maps directly to agent-to-agent oversight: when one AI agent discovers a vulnerability in another AI system, who decides the finding is valid, who decides to escalate, who decides to remediate. Each is a handoff your framework already describes. Your framework was written domain-agnostically, which means it applies to security without modification. That is conceptual depth no domain-specific security designer can match. The vocabulary convergence between this posting and your published work was the kind of outreach hook no other candidate could have claimed.
- Verdict: Ignore-until-trigger. The posting appears pulled from HackerOne's active Ashby feed. At 49 days it was already past the hot window. This is the most significant loss in the scan. Role rubric scored 15/15.
HackerOne — set a weekly monitor
Check HackerOne's careers page and Ashby feed weekly. If the role reappears or a successor posting surfaces, treat it as Act immediately. Do not wait for the next scan cycle.
- Eltropy. AI-powered engagement platform for credit unions and community banks covering voice, text, secure chat, video banking, co-browsing, and agentic AI for financial workflows across 700+ community financial institutions.
- Stage / Size / Capital: Series B. $21M total raised. 51–200 employees.
- Offices: Santa Clara HQ. Role is onsite Mon–Thu in Santa Clara, with a required application question confirming onsite availability. Bay Area-accessible but four days onsite is a significant lifestyle constraint worth naming.
- Value chain: Downstream financial services application. Sits at the member-interaction layer for community financial institutions that lack engineering resources to build their own digital engagement. Design scope is moderate: specific vertical, specific regulatory constraints, limited market ceiling.
- Best at: Agentic AI applied to a trust-sensitive, underserved market. Credit union members and loan officers making financial decisions with AI assistance is a genuine governance design problem, though the scale and complexity sit a tier below the other companies in this scan.
- Signals: Director, Design first published Mar 24, 2026, updated Jul 14, 2026. 122 days old. The Jul 14 revision is the tell: the first search failed. I cannot confirm what specifically changed in the revision from public sources — the Greenhouse API gives me timestamps but not a diff. The pattern is readable: 122 days, a revision at day 112, and the role still open means the original candidate pool didn't match and the committee adjusted. If this role refreshes again or a successor posting appears, the revised language will be the real brief. Comp: $200K–$325K. Leads a 5-designer team in India.
- Tier: Growth-stage platform.
- Rubric scores: Company 8/12 — AI centrality 2, Stage/equity 2 (Series B, $21M, small), Design influence 2, Trajectory 2. Role 11/15 — Total comp 3 ($200K–$325K), Scope 2 (niche market, ceiling likely), Craft depth 2 (leading remote team, some hands-on), AI exposure 2 (financial AI workflows, not cutting-edge), Portfolio value 2.
- Portfolio match: Lead with Alibaba B2B platform, specifically the trust architecture for high-value transactions where both parties need confidence in system recommendations. $50B+ GMV means you have designed trust at financial scale that dwarfs credit union transaction volumes. Alibaba's cross-cultural B2B trust challenge (US buyers, Chinese suppliers) is structurally harder than single-market financial services trust.
- Verdict: Ignore-until-trigger. 122 days old. The Jul 14 revision signals the search is technically still breathing, but the age and 4-day onsite constraint make this low-priority. Trigger: role reposts with hybrid/remote flexibility, or a fresh design leadership posting appears.
Contextual Intelligence
AI augments human decisions in these products, but the consequence surface is thinner. A voice transcription error does not carry the same weight as a missed clinical finding or a false security vulnerability. Real design problems. Lower decision-rights density. Lower portfolio value and scope expandability as a result. Both postings have decayed past their windows.
-
Deepgram. Real-time APIs for speech-to-text, text-to-speech, and production-grade voice agents at scale, serving 200,000+ developers and 1,300+ organizations.
- Stage / Size / Capital: Series C ($130M, Jan 2026, led by AVP). $1.3B valuation. $234M total raised. 51–200 employees.
- Offices: San Francisco HQ. Role is remote, PST preferred.
- Value chain: Upstream infrastructure / API layer. Sits beneath the applications that use voice AI. Design scope concentrates on developer experience, observability tooling, and agent interfaces. Real value, but the design surface is developer-facing, not end-user-facing.
- Best at: Developer experience for voice AI infrastructure. The posting explicitly names "developer workflows, evaluator journeys, design-system evolution, AI observability, agent interfaces, and broader practice transformation over time." That last phrase is function-building language disguised as a Staff IC posting.
- Signals: Staff Product Designer (PST) posted May 12, 2026. 73 days old. Still listed as active in the Ashby feed but well past the 43-day threshold. No direct reports initially but "a natural path into people leadership as the team grows." Series C closed Jan 2026, so the company has been in post-funding build for six months. Slow-fill for a specialized role, or deprioritized. I cannot tell from public signal alone.
- Tier: AI-native.
- Rubric scores: Company 12/12 — AI centrality 3, Stage/equity 3, Design influence 3 (the "practice transformation" language means this hire IS the design influence ceiling — the first design hire who builds the practice defines the ceiling), Trajectory 3. Role 13/15 — Total comp 2 (Staff level), Scope 3 (explicit path to people leadership, practice transformation), Craft depth 3 (developer workflows, AI observability, agent interfaces), AI exposure 3 (voice AI, agent interfaces), Portfolio value 2 (developer tools, less directly publishable).
- Portfolio match: Lead with Alibaba platform redesign, specifically the challenge of redesigning complex professional workflows (B2B search, PDP, homepage) while simultaneously building the design function from a structural gap. Deepgram's hardest design problem is the developer's evaluation moment: a developer integrating voice AI needs to understand whether the model's output is accurate enough for their use case, and the observability tooling must make that judgment fast and confident. The "practice transformation" language maps to your function-building experience at Alibaba, where you named the gap, built the mandate, and led 3 cross-functional sprints. Deepgram needs someone who can do the IC craft work AND build the practice. Your Alibaba case proves both in a single story.
- Verdict: Ignore-until-trigger. 73 days old. Company rubric is strong (12/12) and role rubric clears (13/15), but posting age disqualifies without evidence the search is still active. This is the highest-scoring company in the scan on fundamentals. Trigger: posting refreshes or a new design role appears. If it refreshes, re-evaluate as Act immediately. The function-building scope is real and the company is building from zero.
-
Oura. Health wearable and membership platform where the Oura Ring collects biometric data and the software layer delivers health insights, with emerging AI-powered contextual and agentic interaction models on the product roadmap.
- Stage / Size / Capital: Pre-IPO. Confidential S-1 filing May 21, 2026. Over $900M raised (Oct 2025, led by Fidelity, ~$11B valuation). 501–1,000 employees.
- Offices: San Francisco, Oulu, Helsinki. Role was listed as remote, USA. Comp: $226K–$266K (Region 1).
- Value chain: Vertically integrated consumer health. Oura owns hardware, data pipeline, and software. This Director role sits on the commerce surface (ouraring.com, B2B/partnership fulfillment, membership touchpoints), not the clinical or health-insights layer. That distinction matters for AI exposure and portfolio value.
- Best at: Consumer health brand with a hardware-software loop. The commerce surface is important to the business but is not where the hardest design problems live. "AI-powered contextual and agentic interaction models" appears in the posting as an exploration responsibility, not a core mandate.
- Signals: Director of Product Design, Commerce first seen ~Apr 14–15, 2026. ~100 days old. Greenhouse URL now redirects to an error page and the role does not appear on Oura's visible board. LinkedIn Jobs still lists it but LinkedIn is a lagging indicator. The S-1 filing is interesting context for the equity window, but the commerce focus limits AI exposure and portfolio value.
- Tier: Enterprise platform.
- Rubric scores: Company 9/12 — AI centrality 1 (AI is a feature, not core; commerce role is further removed), Stage/equity 3 (pre-IPO, S-1 filed, $11B valuation), Design influence 2 (Director but commerce-scoped), Trajectory 3. Role 9/15 — Total comp 2 ($226K–$266K), Scope 2 (commerce-focused, ceiling built in), Craft depth 2 (commerce/e-commerce surfaces), AI exposure 1 (AI-contextual mentioned as exploration), Portfolio value 2.
- Portfolio match: If this role were live, lead with Allē, specifically the dual-surface commerce redesign driving 3.2× redemption and $42 CAC across 30M members. The commerce optimization and membership touchpoint work maps directly. Allē's dual-surface challenge (provider-facing + consumer-facing) mirrors Oura's B2B partnership + DTC commerce structure.
- Verdict: Ignore-until-trigger. Greenhouse link is dead. ~100 days old. Commerce focus drops AI exposure to 1/3 and role score to 9/15, below the Act threshold even if the posting were live. Trigger: a health-product or AI-focused design leadership role at Oura. The pre-IPO equity window makes the company worth monitoring, but only for a role on the health-insights surface.
Things to follow up on...
- Brex Staff IC signal: Brex's Staff Product Designer, AI role explicitly names delegation, supervision, review, correction, and trust in automation that touches money, making it one of the strongest decision-rights-density matches among recognized targets this cycle.
- Adobe's AI design cluster: Adobe Design's careers page shows multiple GenAI, Firefly, and GenStudio roles including a Director of Product Design for agentic creative workflows in SF, worth monitoring as a named-target enterprise signal.
- Deepgram's practice-building tell: The Staff IC posting's explicit mention of "broader practice transformation over time" is function-building language that makes a refreshed posting worth treating as Act immediately given the company's 12/12 fundamentals score.
- Oura's IPO equity window: Oura's confidential S-1 filing in May 2026 at an ~$11B valuation means any health-insights or AI-focused design leadership role there would score 3/3 on stage/equity immediately.