Trigger
VP of Design posting, backfill with title upgrade. Posted June 16, updated July 8. Now 38 days old. The departing design leader confirmed publicly that he wrote the original JD and called it "Head of Design." The live posting says VP of Design. Confidence: high on backfill, high on title upgrade.
Verdict: Act.
| AI centrality | Stage & equity | Design ceiling | Trajectory | Total | |
|---|---|---|---|---|---|
| Company | 2 | 2 | 3 | 3 | 10 |
| Comp | Scope | Craft depth | AI exposure | Portfolio value | Total | |
|---|---|---|---|---|---|---|
| Role | 3 | 3 | 2 | 3 | 3 | 14 |
Window
The July 8 update means the first candidate wave didn't close. The title upgrade buys you a narrow reopening, but 38 days on a VP search at a company this well-funded means a second wave is already forming.
Head of Design became VP of Design. That tells you what the committee learned: they need more organizational authority than the departing leader carried. The recalibration opens a brief window. You are precisely on time if you move now. You are late if you wait.
What the new VP inherits
The departing leader's farewell post names what shipped under his watch: ŌURA partnership, cycle tracker, Maven's first AI features with NVIDIA, marketing relaunch. He also credits a colleague with helping "crash land Claude into Maven." Note the verb. Claude was integrated fast, without deep design infrastructure around it.
What the new VP walks into: Maven Intelligence, an agentic orchestration layer launched March 2026. Maven Assistant, a multi-agent system now at 100% of users with explicit human handoffs, including automatic transfer for self-harm language. And a benefits-answering capability that was deliberately withheld because wrong answers could cost members tens of thousands of dollars.
That withheld capability is the commitment point. The design leader who takes this seat inherits live AI surfaces where a bad handoff means financial ruin for a patient. The posting says it plainly: "clarity about what AI enables versus where human judgment matters most." The AI features already exist, are live, are at full deployment. The mandate is designing the trust architecture around them.
Your angle
Narrative fit: You've built exception-first systems in pharma where catching failures upstream saved $20M+ annually and catching them at the delivery gate cost 3–5× more. Maven's benefits-answering problem is the same structural pattern in healthcare. The VP who wins this role will be the one who can name that pattern.
Lead with Thermo Fisher. Support with Red Cross. Agentic Labs for AI credibility.
Thermo Fisher is the direct analog. Regulated industry. Exception-first design. The core insight that exceptions surfacing at the delivery gate cost 3–5× more than if caught upstream. Your mySupply work ($20M+ margin recovered, 6 pharma partners, 100% adoption) proves you've designed these systems where the human gate is non-negotiable.
Red Cross reinforces the mission-critical dimension: $847K disbursed, 6 systems consolidated to 1, national deployment under real constraints. Agentic Labs, specifically Carrier IQ in regulated InsurTech, and the Trust essay's Decision Gate handoff map directly to Maven's clinician-review trigger when elevated risk signals surface.
One additional match if they probe platform complexity: Allē's dual-surface model (30M consumer members + 40K provider practices) mirrors Maven's B2B2C structure. Don't lead with it. Deploy only if asked.
Competitive landscape
Healthcare AI design leader (moderate confidence): candidates from Oscar Health, Ro, Hims, or Cityblock with direct healthcare AI experience. They outperform you on domain familiarity. They likely underperform on regulated 0→1 build proof and agentic trust architecture vocabulary. Your counter: your exception-first framework is transferable and named. Theirs is probably implicit.
Enterprise design director (moderate confidence): someone from a large health system's digital product team, or health-tech enterprise like Epic or Cerner. They outperform on healthcare domain and possibly team scale. They underperform on AI-native credibility and speed. Your counter: TinyFish velocity (3 products, 3 months) and Agentic Labs prove you build in the medium, not around it.
AI-native design leader (speculative): a candidate from Anthropic, OpenAI, or a well-funded AI startup. They outperform on AI brand recognition. They almost certainly underperform on regulated-industry proof and mission-critical deployment experience. Your counter: Thermo Fisher + Red Cross. Nobody from an AI startup has $20M in recovered margin from exception-first design in pharma.
You win at the intersection. No other candidate is likely to carry regulated 0→1 build proof, a published agentic trust framework, and current production AI experience simultaneously.
Who to contact
Primary: Jason Lee. Listed as CPO on the posting and Maven's author page. LinkedIn.
Maven's Onward 2026 event page and The Org both list Lee as COO, with Jaya Savkar as SVP, Head of Product. The posting still names the CPO. Address outreach to Lee as the design sponsor regardless. Verify the actual reporting structure in the first conversation.
The departing leader publicly credits Lee with giving design "real room to work."
Public fingerprint: Thin. Lee's most recent LinkedIn activity (~3 weeks ago) shared Maven's nationwide consumer launch: "Everyone deserves access to quality care." A 2023 Q&A has him naming patient engagement as "one of healthcare's biggest product problems." A 2021 interview says he wants to "enable providers to operate at the top of their licenses." No conference talks or bylined articles in the last six months.
Secondary: Jaya Savkar, SVP, Head of Product per the Onward 2026 page. If Lee's title has shifted to COO, Savkar may be the actual hiring manager. Her public fingerprint is thin from available sources. Search her LinkedIn before sending. If you find substantive recent posts on design or product, send a parallel message with the same angle. If her signal is as thin as Lee's, route through Lee first and let the conversation surface the real reporting line.
Warm path: Cold on available evidence. No direct network overlap surfaced. Semi-warm route possible through the Maven design community: the departing leader (Eli Horne), Loric Avanessian (product design manager, featured in Figma's coverage of Maven's AI prototyping work), or Kalliope White (named in Horne's departure post). Check your LinkedIn graph for shared connections before defaulting to cold.
Channel: LinkedIn DM to Lee.
Outreach messages
Warm message (if a connection to the Maven design team surfaces):
"Maven deliberately withheld benefits answering from Assistant until the trust threshold was right, because a wrong answer could cost a member tens of thousands of dollars. That decision is the design problem I keep returning to: where does the system stop and the human decide? I designed exception-first systems at Thermo Fisher that caught $20M in pharma supply chain failures upstream instead of at the delivery gate, and I've published a framework for agentic trust handoffs at junochen.com/trust-is-the-new-interface. I'd love 20 minutes on how Maven is thinking about clinician handoff surfaces as Intelligence scales. [Mutual connection] thought this conversation would be worth both our time."
Cold message (to Lee):
"You said in your 2021 Maven interview that you want to enable providers to operate at the top of their licenses. With Assistant now at 100% of members and an agentic layer mediating the care, that ambition has a new design problem underneath it: where does the AI stop and the clinician decide, and what does that boundary look like to a member trusting the system with their health? I designed exception-first systems in pharma at Thermo Fisher ($20M+ margin recovered, 100% partner adoption) and published a trust-handoff framework for agentic systems at junochen.com/trust-is-the-new-interface. I'd welcome 20 minutes on how you're thinking about the VP of Design mandate as Maven Intelligence scales."
Subject line: The handoff surface between Maven Assistant and a clinician
What not to say:
- Do not reference the Lee title ambiguity (CPO vs. COO). Let him self-identify. Raising it signals you've been reading org-chart aggregators, which feels invasive rather than prepared.
- Do not call this a "Head of Design" role. The departing leader used that title. The posting upgraded it. Using the old title signals you found the departure post but didn't read the actual listing.
Cover letter draft
Maven deliberately withheld benefits answering from Assistant because a wrong answer could cost a member tens of thousands of dollars. That is the design problem I've spent five years solving across regulated industries, and it is the reason I'm writing.
At Thermo Fisher, I built mySupply from zero: a digital supply chain platform for six pharma partners across three continents where exceptions caught upstream saved $20M+ annually and exceptions caught at the delivery gate cost 3–5× more. The design insight that transfers directly: in regulated environments, the most important interface is never the happy path. It is the moment the system surfaces a risk signal and a human has to decide whether to act. At the American Red Cross, I consolidated six legacy systems into one national disaster-relief platform that disbursed $847K in its first two weeks. Every workflow carried FEMA compliance constraints. Every error had real human cost.
I'm currently Head of Product at TinyFish, an enterprise web agent platform (Series A), where I shipped three products in three months and work daily with agent traces, auditability, and governance challenges in production. I moved to product to build AI-natively from zero. I'm returning to design because I want to apply that technical depth to a specific, high-stakes vertical. Healthcare AI trust design is the problem I'm most compelled by.
My published portfolio and trust-handoff framework are at junochen.com. I'd welcome the chance to discuss how Maven is thinking about trust architecture as Intelligence and Assistant scale.
Resume emphasis guidance
Craft depth: 2 (leads 15, strategic + hands-on). AI exposure quality: 3 (conversational AI handoffs, agent workflows, clinician trust boundaries).
- Lead with Thermo Fisher mySupply. Regulated 0→1, exception-first design, pharma compliance. Expand the exception-upstream insight. It maps directly to Maven's guardrail design problem.
- Follow with Red Cross. Mission-critical, national deployment, multi-system consolidation. Reinforces high-stakes trust design at scale.
- Third: Agentic Labs (Carrier IQ specifically). Regulated industry + agentic automation. Demonstrates all five trust handoffs.
- Fourth: Allē. Dual-surface (consumer + provider) maps to Maven's B2B2C model. Keep visible but don't lead.
- De-emphasize Alibaba. Enterprise B2B platform scale is relevant but not the lead story here. Move below the fold.
- TinyFish in header: "Head of Product at TinyFish (enterprise web agent platform, Series A)." One line: shipped 3 products in 3 months, daily work with agent traces and governance in production.
The non-obvious thing
The departing leader called the role "Head of Design" in his farewell post. He said he wrote the JD. The live posting says "VP of Design." That upgrade almost certainly happened during the July 8 revision, after the first candidate wave was evaluated.
The committee concluded they need more organizational authority than the departing leader carried. The team of 15 is already built. The AI surfaces are already deployed. But the design authority to shape how those surfaces evolve as Maven Intelligence scales from orchestration layer to clinical decision infrastructure does not yet exist. Candidates who pitch operational continuity will miss this entirely. Pitch the expanded mandate.
Warning signs
Leadership title instability. Lee's title appears to have shifted from CPO to COO, with Savkar taking SVP Head of Product. If the reporting line has changed and the posting hasn't been updated, you may be interviewing with a different buyer than the one described. Verify in the first conversation.
Glassdoor signal. 3.2 out of 5 from 236 reviews. Senior management scores 3.0. A June 2026 review cites "constant leadership changes." Company-wide, not design-specific, but consistent enough across multiple reviews to flag. Ask about design team tenure and how many of the 15 reports have been there longer than a year.
Neither is disqualifying. Both are verify-in-conversation items.
Post-outreach signals
Maven is a growth-stage healthcare platform. Expect slightly slower response cycles than pure tech.
- Response within 3–5 business days: Normal. Healthcare companies carry compliance-adjacent process even in recruiting.
- Silence past 7 business days: The search may be further along than the posting suggests. One follow-up at day 8 referencing Maven's Onward 2026 event (September 30), which has an agenda item on AI guardrails. Natural hook.
- Recruiter screen scheduled within first contact: Fast-moving. Prioritize prep immediately. Lead the portfolio walkthrough with Thermo Fisher's exception-first design, not Alibaba.
- Calendar availability in the first reply: Move Maven to the top of the daily queue. This search is 38 days old and they want to close.
- No reply after one follow-up: Close the loop. Monitor for a posting refresh or new design leadership signal on LinkedIn.
- Maven's Onward 2026 agenda: The September 30 event includes a session called "AI for Families: Tactical Strategies for Guardrails that Accelerate Innovation" with Lee, Savkar, and CMO Neel Shah, which will surface the exact guardrail language Maven's leadership uses internally.
- Maven Assistant at full rollout: Independent coverage reports that Maven Assistant expanded from 20% to 100% of users and shipped benefits answering after initially withholding it, so watch for public performance data or design retrospectives that reveal what the trust threshold looked like in practice.
- Loric Avanessian's Figma Make work: Maven's product design manager was featured describing how AI prototyping shipped an MVP in fewer than four sprints, which tells you the team is already building with AI tools and the VP will inherit that velocity expectation.
- Lee's title resolution: Maven's author page still says CPO, but The Org now lists Lee as COO with Savkar as SVP Head of Product, so monitor LinkedIn for a formal title update that would confirm the reporting-line shift before your first conversation.