Scan window: September 4, 2026. Act on Within and Stripe this weekend. OpenAI by Wednesday September 10. Gray Swan through September 17.
Four packets below, ranked by when the outreach value degrades. Within's posting hits day 22 tomorrow, Stripe's on Sunday — both lose hot-window advantage Monday morning. OpenAI crosses September 10. Gray Swan has room, but the mandate overlap is close enough that I would not sit on it.
Workiva's application closes September 7. Submit it, but keep it in its own lane. It is a form, and it should not eat a slot that belongs to one of these four.
Upgrade from last cycle: Within now has a confirmed design buyer, Tara Ting, Head of Product Design, replacing the Nithya Subramaniam route. Katie Dill and Ian Silber both gave 2026 interviews specific enough to anchor outreach language. Every message below borrows words the recipient said in public.
Act Now: Within
What changed: Tara Ting is publicly listed as Head of Product Design. The Staff Product Designer posting appeared August 14. Today is day 21. Tomorrow it enters the 22–42-day Watch band.
Why it matters: Within builds an AI-native platform that captures how organizations work — workflows, decisions, handoffs, judgment calls — and turns that into a context graph agents can use. Their public language says the system is "built to map work, not track workers," but the product pages never show how the stated review and anonymization controls behave in the interface. That gap between stated trust principles and shipped trust surfaces is the problem your Trust essay names. Lead with the Agentic AI pillar.
Confidence: High on Ting's role and posting age (LinkedIn, Ashby ATS, MIT Orbit corroboration). Moderate on mandate match — inferred from product architecture, not confirmed by Ting.
Window: Day 21 of 21. Crosses to Watch tomorrow.
Evidence ceiling: Trust essay, Agentic Labs production work, TinyFish as role context. Alibaba is secondary here. This is an AI-native IC role, not an enterprise leadership mandate.
Scores: Company 10 (AI centrality 3, Stage/equity 2, Design influence 3, Trajectory 3). Role 14 (Comp 2, Scope 3, Craft 3, AI exposure 3, Portfolio value 3). Verdict: Act.
Two of those numbers need a footnote. Stage/equity scores 2 rather than 3 because Within's funding stage is not publicly confirmed in any source I checked. Product maturity and team-building activity suggest mid-stage, but the equity read is speculative until you verify the round. Design influence scores 3 on the strength of Ting's public framing that EPD "defines the business rather than supports it" — moderate confidence, since no org-structure signal corroborates it. Both could move after a first conversation.
Tier archetype: AI-native. Tier Playbook: Read the AI-native playbook before outreach. Battlecard: Flag as candidate.
Move
(a) Who: Tara Ting, Head of Product Design. She was not publicly identified in this role last cycle, which means she is likely building her team right now. Her LinkedIn frames EPD as functions that define the business rather than support it, and she describes the ideal hire as someone who "raises the ceiling."
Contact ladder: Ting (design buyer) → Nischal Nadhamuni, co-founder (mandate sponsor, appears in company events) → Nithya Subramaniam, Senior Product Designer (routing contact from the prior cycle — if you engaged with Nithya then, that is an internal path to Ting).
(b) Warm path: Cold. If prior outreach to Nithya generated any response, use it.
(c) Outreach plan: (1) Engage with Ting's LinkedIn content; she posts about hiring and product. (2) Connection request with the note below. (3) Full message after acceptance. (4) Apply directly today. (5) Follow up in five business days if no response.
Internal reason to act now: The posting crosses out of the hot window tomorrow. External reason the message feels timely: Ting is actively promoting the role and building her team. She expects strong inbound this week.
Intended easy response: "Thanks for reaching out — let's find 20 minutes next week."
(d) Connection request (~270 chars):
"Hi Tara — your framing of EPD defining the business, not supporting it, maps to how I approach design for AI-native enterprise products. I've been writing about trust design for agentic systems (junochen.com) and would value learning about what you're building at Within."
(e) Full outreach message (~250 words):
"Hi Tara,
I saw the Staff Product Designer posting and wanted to reach out directly because the mandate maps closely to what I've been building toward.
Within's architecture — capturing how organizations actually work, then giving agents that context — creates a design problem I've been thinking about publicly: how do you design trust into systems where an AI agent is observing, interpreting, and acting on human work patterns? The interface has to make the system's understanding visible and correctable, or adoption stalls at the people who need it most.
I wrote about this in 'Trust Is the New Interface' (junochen.com) — a framework for the five handoffs where human-agent trust is built or broken: intent-setting, in-progress monitoring, output review, decision gates, and feedback loops. The framework came from shipping an enterprise AI agent platform as Head of Product at TinyFish, where I took the product from zero to production in three months.
The design proof is at junochen.com — three production AI products (Brand Pulse, Retail Velocity, Carrier IQ through Agentic Labs), plus the Alibaba.com B2B platform redesign where I led design and research for North America ($50B+ GMV, +20% daily transactions). The enterprise trust problem at scale is something I've shipped against.
I'd like to hear how you're thinking about the trust layer in Within's product — particularly how Companion's observation model earns user confidence in practice. Would 20 minutes work?
Best, Juno"
(f) What not to say: Do not call Within's product "automation" or "RPA." Their positioning is about understanding work. Do not reference the old name Klarity. Do not lead with management experience — this is a Staff IC role and Ting wants builders.
Interview Prep
- Format: Portfolio review weighted toward hands-on craft. Within values velocity; a company event described using AI to compress concept-to-prototype cycles. Show speed and iteration.
- Portfolio lead: Trust essay framework mapped to Within's surfaces. Companion observation maps to intent-setting; Company Brain maps to output review; Advisor recommendations map to decision gates. Back it with Agentic Labs production work.
- Key question: "How does Companion handle the moment when an employee disagrees with how their workflow has been captured — what does correction look like in the interface?"
- Competitive differentiator: You have a published trust framework, production agentic AI work, and enterprise scale proof. Most candidates at this level bring one of the three.
Act Now: Stripe Data & AI
What changed: Katie Dill's Dive Club interview hands you her evaluation language directly. The Senior Staff Product Designer, Data & AI posting appeared August 15. Day 20. Enters Watch Sunday.
Why it matters: Senior Staff IC role designing how businesses understand and act on their data through AI-generated intelligence. The prior cycle routed through Eeke de Milliano because the design-side manager was unresolved. Dill is the functional design leader and the stronger senior route. Her stated criteria — curiosity, agency, comfort with ambiguity, prototyping over pitch decks — describe how you already work. Her observation that AI has made a competent "seven out of ten" cheap, pushing differentiation toward taste and systems thinking, is your opening. Bridge the Agentic AI and Enterprise pillars.
Confidence: High on posting age and Dill's role (Stripe careers, Tech:NYC corroboration). Moderate on Dill as decision-maker for this specific req; she runs design org-wide, and the Data & AI team's direct design manager is not publicly identified.
Window: Day 20 of 21. Crosses to Watch Sunday.
Evidence ceiling: Alibaba enterprise scale leads. Stripe's evaluation pattern rewards systems thinking across a large, expanding product surface. Agentic Labs and the Trust essay are the AI bridge. TinyFish is role context only.
Scores: Company 9 (AI centrality 2, Stage/equity 2, Design influence 3, Trajectory 2). Role 13 (Comp 3, Scope 2, Craft 3, AI exposure 2, Portfolio value 3). Verdict: Act.
Tier archetype: Enterprise platform. Tier Playbook: Read the Enterprise platform playbook. Battlecard: Flag as candidate. Prior cycle coverage exists, but no full Battlecard.
Move
(a) Who: Katie Dill, Head of Design, Stripe. Her stated criteria: curiosity about AI and the changing design practice, agency to act on it, executable prototyping (she told designers to stop preparing pitch decks and put working prototypes in front of collaborators), systems awareness across Stripe's expanding ecosystem, high craft bar.
Contact ladder: Dill (senior design route) → Eeke de Milliano (Data & AI product leader, identified prior cycle) → apply directly.
(b) Warm path: Cold to Dill. If prior outreach to de Milliano generated a response, reference that conversation.
(c) Outreach plan: (1) Engage with Dill's LinkedIn content; she recently shared a hiring post for Agentic Commerce. (2) Connection request below. (3) Full message after acceptance. (4) Apply directly today. (5) Follow up in five business days.
Internal reason: The posting crosses out of the hot window Sunday. External reason: Dill is actively promoting design roles and expects strong candidates to surface now.
Intended easy response: "Thanks — let me connect you with the Data & AI design lead."
(d) Connection request (~280 chars):
"Hi Katie — your point about AI making a competent 7/10 inexpensive and shifting the work to taste and systems maps to what I've been designing at the intersection of enterprise data and AI intelligence. I'd welcome a conversation about the Data & AI challenge. junochen.com"
(e) Full outreach message (~260 words):
"Hi Katie,
I'm reaching out about the Senior Staff Product Designer role on the Data & AI team. Your Dive Club interview framed something I've been working through in practice: when AI can produce a competent baseline, the design work that matters is systems thinking, taste, and the judgment calls that hold a product together across surfaces.
That's the problem I've been solving in two directions. At Alibaba.com, I led design and research for the B2B platform redesign across North America — $50B+ GMV, three cross-functional sprints across homepage, search, and product detail pages, +20% daily transactions. The challenge was making an expanding product ecosystem coherent without requiring users to absorb everything at once — which sounds close to the problem you described for Stripe's growing product surface.
On the AI side, I shipped three production AI products through Agentic Labs (Brand Pulse, Retail Velocity, Carrier IQ — live at junochen.com) and wrote 'Trust Is the New Interface,' a framework for designing the handoffs where humans and AI agents calibrate trust. Most recently as Head of Product at TinyFish, I took an enterprise AI agent platform from zero to production in three months.
I prototype to think, not to present — which I gather is how your team works. Could we find 20 minutes to talk about how the Data & AI team handles the trust problem when AI-generated insights drive financial decisions?
Best, Juno"
(f) What not to say: Do not lead with a polished case study narrative; Dill values prototyping over decks. Do not position yourself as "an AI designer" generically — her criteria are curiosity and agency applied to AI, not an AI pedigree. Do not reference the Agentic Commerce role. That is Dan Nelson's team.
Interview Prep
- Format: Portfolio review where they want to see how you think through problems rather than what you finished. Dill described close critique paired with permission to take risks. Stripe uses Protodash for interactive prototyping, so comfort with a prototype-driven culture matters.
- Portfolio lead: Alibaba enterprise scale plus Agentic Labs AI production. The combination is the differentiator — Senior Staff candidates at Stripe typically have deep enterprise experience or AI-native product experience, rarely both with published work behind each.
- Key question: "How does the Data & AI team handle the moment when AI-generated intelligence contradicts what a business owner believes about their own data?"
Act Now: OpenAI Growth–Codex
What changed: Ian Silber gave two interviews — Lenny's Podcast and Dive Club — that expose concrete evaluation criteria. The Product Design Lead, Growth–Codex posting appeared August 19. Day 16. Enters Watch September 10.
Why it matters: Highest comp ceiling on the board, and a player-coach role at the company defining the category. Silber's criteria are specific: designers who stay close to the model, test what it does well, watch where it breaks, and ask whether a problem can be solved with tokens or model behavior before adding interface. His "just do less" systems principle — check whether an existing component can carry the experience before building a new one — is a close cousin of your Trust essay's argument about when to let the agent act and when to insert a human checkpoint. Lead with the Agentic AI pillar.
Confidence: High on posting age and Silber's philosophy (Ashby ATS, Omega VP corroboration, two published interviews). Moderate on Silber as direct hiring manager for this req. High on interview format (OpenAI guide: 4–6 hours, 4–6 people, 1–2 days).
Window: Day 16 of 21. Five days before the Watch band.
Evidence ceiling: Trust essay plus Agentic Labs production work, with TinyFish as bridge narrative. Alibaba is secondary; this audience cares about model-proximity, not enterprise scale for its own sake.
Scores: Company 12 (AI centrality 3, Stage/equity 3, Design influence 2, Trajectory 3). Role 15 (Comp 3, Scope 3, Craft 3, AI exposure 3, Portfolio value 3). Verdict: Act.
Design influence scores 2, not 3. OpenAI is research-led. Design executes with real influence but does not set product direction. Know that before you walk in, and position as a collaborator with the research culture rather than someone who will assert design authority over it.
Tier archetype: AI-native. Tier Playbook: Read the AI-native playbook. Battlecard: Flag as candidate.
Move
(a) Who: Ian Silber, Head of Product Design, OpenAI. He does not require prior AI company experience and values curiosity, adaptability, and evidence of hands-on experimentation with AI systems. He prefers complementary specialists over generalists who claim every skill.
Contact ladder: Silber (functional design leader) → Growth–Codex product lead (not publicly identified; that is a gap) → apply directly.
(b) Warm path: Cold.
(c) Outreach plan: (1) Engage with Silber's content if he shares the podcast appearances. (2) Connection request below. (3) Full message after acceptance. (4) Apply directly today. (5) Follow up in five business days.
Internal reason: Highest-value role on the board by comp and by portfolio value. External reason: Silber has spent two major podcasts describing what he looks for in designers. He is signaling openness to inbound.
Intended easy response: "Thanks — applying through the posting is the best path, but I'll flag your profile to the Growth team."
(d) Connection request (~265 chars):
"Hi Ian — your framing of solving problems with tokens before pixels and 'just do less' as systems advice aligns with how I've been thinking about trust design for agentic systems. I'd welcome a conversation about the Growth-Codex design challenge. junochen.com"
(e) Full outreach message (~270 words):
"Hi Ian,
I'm reaching out about the Product Design Lead role for Growth–Codex. Your Dive Club interview articulated something I've been working through in practice: the question of whether a problem should be solved with model behavior or interface, and the discipline of examining whether an existing component can carry the experience before adding a new one.
I wrote about a version of this in 'Trust Is the New Interface' (junochen.com) — a framework for the five handoffs where human-agent trust is built or broken. The argument is that most of what we call 'AI UX' is actually a trust-calibration problem: the interface needs to help users develop accurate mental models of what the agent can and cannot do, then get out of the way. That is a 'just do less' argument applied to the trust layer.
The production proof: three AI products shipped through Agentic Labs (Brand Pulse, Retail Velocity, Carrier IQ — live at junochen.com), each solving a different trust-calibration problem in agentic workflows. Most recently as Head of Product at TinyFish, I took an enterprise AI agent platform from zero to production in three months — staying close to the model, testing where it broke, designing around its actual capabilities.
My specific depth is in the trust and supervision layer of agentic systems — where the human decides whether to watch, verify, or delegate. I'd like to discuss how Growth–Codex is approaching that calibration problem as Codex's capabilities expand. Would a short conversation make sense?
Best, Juno"
(f) What not to say: Do not lead with prior AI company experience as the credential; Silber said explicitly that it is not required and that he values curiosity over pedigree. Do not present a polished deck — OpenAI favors prototypes and videos. Do not claim systems-design expertise abstractly. Show it through the "just do less" lens he uses.
Interview Prep
- Format: 4–6 hours, 4–6 people, 1–2 days. Evaluated on domain expertise or ramp ability, collaboration, communication, openness to feedback, mission alignment. Expect to demonstrate systems thinking through a specific design problem.
- Portfolio lead: Trust essay plus Agentic Labs, framed through Silber's "model as the product" lens. Show how you designed around model capabilities and limitations rather than around a fixed spec.
- Key question: "How does the Growth–Codex team balance experimental releases with the parts of the experience that need cohesion — where is the current tension?"
- Leadership dynamic: Silber wants complementary specialists. Position as the trust and supervision specialist. Show curiosity about model capabilities rather than authority over them.
- Competitive differentiator: The published trust framework with production proof behind it. Plenty of candidates will bring AI curiosity or enterprise shipping experience; your edge is that the Trust essay names the exact problem Silber described valuing, and the Agentic Labs work shows you've already solved versions of it.
Act Soon: Gray Swan AI
What changed: Meredith McDermott (Head of User Experience, previously LinkedIn and Duolingo, CMU MDes 2015) has public statements that give you a precise angle. The Staff Product Designer posting appeared August 27. Day 8. Hot window through September 17.
Why it matters: Gray Swan builds Arena, an adversarial research environment where incentivized human red-teamers probe AI models for vulnerabilities before release. Two user populations: red-team participants (challenges, scoring, leaderboards, reputation) and model builders (threat intelligence, pre-deployment testing). NIST has partnered with Gray Swan on large-scale agent red-teaming competitions. McDermott's framing — "At Duolingo, we made products delightful. At Gray Swan, we make them defensible" — tells you how she thinks about the design problem. Her team's published position that safety is a design constraint from discovery onward, and that teams should define what "good," "safe," and "useful" mean for the use case with an explicit rubric, sits right on top of your five-handoffs framework. Lead with the Agentic AI pillar.
Confidence: High on posting age and McDermott's role (Ashby ATS, Madrona corroboration, CMU event bio, dscout panel). High on mandate match. Moderate on comp; Series A, equity upside speculative.
Window: Day 8 of 21. No immediate decay pressure.
Evidence ceiling: Trust essay leads, Agentic Labs production work backs it. The overlap between your published framework and McDermott's stated design philosophy is the closest on the board, and the message should be built around it.
Scores: Company 9 (AI centrality 3, Stage/equity 1, Design influence 2, Trajectory 3). Role 14 (Comp 2, Scope 3, Craft 3, AI exposure 3, Portfolio value 3). Verdict: Act.
Stage/equity scores 1 because Gray Swan is Series A and the rubric reads that as early-stage execution risk. The NIST partnership and frontier-lab customer base are strong trajectory signals, but they do not change the funding math. Comp floor is uncertain and equity is speculative. The mandate match justifies the outreach — go in knowing the comp conversation may require flexibility.
Tier archetype: AI-native with regulated characteristics (AI safety, NIST partnership). Tier Playbook: Read the AI-native playbook, and pull the Healthcare/Regulated playbook for safety-constraint framing. Battlecard: Flag as candidate.
Move
(a) Who: Meredith McDermott, Head of User Experience, Gray Swan AI. Her dscout panel positioned safety as a source of stronger product ideas. She has also connected red teaming to both product evaluation and participant professional development.
Contact ladder: McDermott (design buyer) → apply directly. No mandate sponsor or routing contact publicly identified.
(b) Warm path: Cold. The CMU design network is a possible indirect path if you have CMU connections.
(c) Outreach plan: (1) Engage with McDermott's LinkedIn content. (2) Connection request below. (3) Full message after acceptance. (4) Apply directly. (5) Follow up in five business days.
Internal reason: Strongest mandate match on the board — the overlap between your Trust essay and their stated design philosophy is near-exact. External reason: The posting is eight days old and McDermott is building her team at a Series A company. She wants strong inbound now.
Intended easy response: "I read the essay — let's talk."
(d) Connection request (~265 chars):
"Hi Meredith — your distinction between delightful and defensible frames the design problem I've been working on: making AI systems trustworthy through interface design, not just guardrails. I'd love to connect about what you're building at Gray Swan. junochen.com"
(e) Full outreach message (~190 words):
"Hi Meredith,
I saw the Staff Product Designer posting and wanted to reach out because your dscout session articulated a position I've been developing in my own work: that safety is a design constraint from discovery onward, not a checkpoint at the end.
I wrote about this in 'Trust Is the New Interface' (junochen.com) — a framework for the five handoffs where human-agent trust is calibrated: intent-setting, monitoring, output review, decision gates, and feedback loops. The framework came from shipping agentic AI products in production — three through Agentic Labs (Brand Pulse, Retail Velocity, Carrier IQ, all at junochen.com) and an enterprise agent platform as Head of Product at TinyFish.
Arena's two-sided design problem — making red-team participation productive and visible while giving model builders actionable threat intelligence — is where trust design and safety design converge. I'd like to hear how your team navigates that convergence in practice.
Best, Juno"
(f) What not to say: Do not frame AI safety as a limitation or a compliance burden. Her team positions early safety work as a source of stronger product ideas. Do not lead with enterprise scale proof; lead with the trust framework and the AI production work.
Interview Prep
- Portfolio lead: Trust essay plus Agentic Labs. Map the five handoffs to Arena's surfaces: challenge discovery (intent-setting), attack attempts (monitoring), submission scoring (output review), vulnerability confirmation (decision gate).
- Key question: "How does Arena handle the gap between what a red-teamer discovers and what a model builder can act on — is that translation a design problem your team owns?"
- Competitive differentiator: The trust framework applied to AI safety, backed by production agentic AI. Where other candidates will bring either UX research depth or AI technical depth, you have a public body of work that names the problem Gray Swan is hiring against and production experience solving adjacent versions of it.
Director of Product Design, Sustainability Management closes Sunday. Submit it. Twenty minutes of form-filling, not a conversation — do not let it displace outreach time for Within, Stripe, or OpenAI this weekend.
Active Lists
Top Match (Act threshold, awaiting activation readiness)
- Vanta — Director of Product Design. The Head of Design search is still advertised with no public fill identified. The Director's reporting line remains the open question. If the Head role fills first, the incoming leader picks their own reports and the Director window either closes or reopens under different criteria. If CPO Jeremy Epling is the direct report, that is a scope opportunity. Missing before activation: confirmation of Head search status. Ask about reporting structure in the first conversation. Tier archetype: Enterprise platform. Tier Playbook: Enterprise platform.
- Anthropic. Same-ID posting resurfaced. Could mean the first search failed and the criteria were revised; could be routine ATS maintenance. Missing: confirmation the refresh reflects a genuine re-opening. Watch for language changes in the posting. Tier archetype: AI-native. Tier Playbook: AI-native.
- Headway — Senior/Staff Product Designer, Provider Experience. Act-level opportunity (11 / 14–15). The posting crosses to Watch September 7. Not activated this cycle because the direct design manager is still unconfirmed. Roland Tiangco is the strongest provider-scope design route; Ellen Dong and Jake Poses are function and product sponsors. Missing before activation: a confirmed hiring manager. Tier archetype: Healthcare/Regulated. Tier Playbook: Healthcare/Regulated.
Watch (named trigger for escalation)
Suno (AI-native), Ramp (Growth-stage platform), Salesforce, Atlassian, Adobe, Rubrik (Enterprise platform) — no new signals this cycle. Triggers: a new design leadership posting at Suno or Rubrik, a design leadership departure signal at Ramp, a Director+ design posting in an AI-adjacent product area at Salesforce, Atlassian, or Adobe.
Deprioritize
- Gusto — Prior posting stale 43+ days. Pause. Revisit on a new design leadership req.
- Amplitude — Prior posting stale 43+ days. Pause. Revisit on a new design leadership req.
- Capital One — Signal decayed. Pause. Revisit on the next design leadership posting.
- Superhuman — Signal decayed. Remove from active monitoring. Revisit only on a funding event or a design leadership departure.
- Brex — Posting stale 43+ days. The Capital One acquisition closed April 7, which removes the pre-IPO equity premise that supported the company score. Remove from active monitoring. Revisit only on a materially new Capital One-era design mandate with confirmed AI scope and compensation.
- Abridge — Posting stale, indexed roughly four months ago. The mandate is still a strong healthcare-AI fit but it is not a timely trigger. Pause. Revisit on a repost, a new Staff or leadership seat, or an identified design leader publicly hiring.
Window Status Changes
Upgraded: Anthropic (Ignore → Watch, same-ID resurfacing).
Downgraded: Gusto, Amplitude, Capital One, Superhuman, Brex, Abridge (all Watch → Deprioritize).
Within (hot → Watch, September 5) · Stripe Data & AI (hot → Watch, September 6) · Headway (hot → Watch, September 7) · OpenAI Growth–Codex (hot → Watch, September 10).
- OpenAI's Identity role: The Product Designer, Identity posting lacks a Staff label but its $245K–$310K range and ownership of agent identity, permissions, and human-to-agent authorization place it above many formally senior roles — worth tracking as a rubric-first exception if the Growth–Codex conversation opens a broader relationship with Silber's org.
- Vanta's Head of Design timeline: Deb Kawamoto's July 9 departure post explicitly called the Head of Design opening her backfill and named the leadership bench she rebuilt — that bench is who the incoming Director would work alongside, and whether the Head search fills first determines whether the Director window stays open.
- Within's product documentation gap: Within's platform pages describe observation as "user-controlled, anonymized, and reviewed" but do not expose the actual interface states for employee consent, correction, or deletion — that gap is your interview question and potentially a portfolio-value surface if the role lets you design the trust layer they describe but haven't shipped visibly.
- Gray Swan's NIST validation: NIST's agent red-teaming research blog confirms Gray Swan as a partner for large-scale competitions testing agent hijacking and indirect prompt injection under adaptive human pressure — useful context for framing the Arena design problem as federally validated, not startup-speculative.