Vanta: The Search Is Five Weeks Old. Move This Week.

Updates the prior Vanta battlecard (company 10/12, role 14/15, Act verdict). New intelligence: Epling's public design philosophy, the director bench's silence, and accelerating product risk from sandbox-enabled agents. Scoring holds. Outreach sharpens.

The Head of Design posting went live July 6. Five weeks in — posting language unchanged, no interim leader announced, no public evidence of a shortlist. At comparable-stage companies, VP Design searches close in roughly one quarter from posting. You have three to five weeks before this pipeline narrows to finalists.

Seven events, eighteen months

July 2023. Vanta hires Jeremy Epling as CPO. Product, engineering, and design consolidated under him from day one. He comes from GitHub (VP Product, Actions/Codespaces/npm) after 16 years at Microsoft — a product executive who credits early managers with teaching him to value design craft, and who saw at GitHub what a high-bar design org could hold.

April–May 2025. Epling goes public with the philosophy that will shape this search. In an interview with Aakash Gupta: every PM gets access to V0, specs become prototypes, designers submit pull requests. His phrase:

Internal mantra: "seeing is believing."

April 2026. Vanta crosses $300M ARR, tripled since 2024. Fortune confirms independently. The $4.15B valuation from the July 2025 Series D anchors the equity picture.

May 2026. Epling writes that Vanta shipped 20 agentic workflows and is targeting 90+ by year-end. The product is moving from recommendations into write operations on risk registers, automated control onboarding, scheduled agents running compliance work daily.

June 2026. Vanta launches Agent for Risk with "human in the loop, always" as the stated principle. Same month, Deb Kawamoto leaves, having grown design from 5 to roughly 45 people.

July 2026. Vanta gives its Agent isolated Linux sandboxes to analyze files, execute scripts, and produce artifacts. Their engineering team writes that sandbox support introduced "new ways for overall quality to degrade" — unnecessary sandbox invocations, partial dataset inspection, authorization boundary crossings. The Head of Design posting goes live. Vanta's design careers page still lists Kawamoto as VP Design.

August 2026 (now). The four directors Kawamoto named — Conor Sheehan, Jerrod Larson, Braden Kowitz, Jenny Chang — all appear to remain at Vanta. None has publicly claimed an interim role or promoted the search. The posting shows no detectable modification since it went live.

What the posting is actually buying

Read the posting language against Epling's public philosophy: this is a search for the first design leader who will operate inside Epling's blended model, not a backfill for the leader who built the org under the previous one.

Kawamoto grew the team from 5 to 45 under a contract where design was a distinct function with its own craft identity. Epling has spent three years building toward something different — designers submitting PRs, PMs building prototypes, discipline boundaries dissolving by intent. The posting's emphasis on AI-native practices, definitions of done, and launch review reads like a quality-governance mandate for a product org that ships faster than any single function can review.

Most candidates in this pipeline will pitch themselves as design leaders who can run a 40-person org. The posting is asking for someone who can hold quality across an org where the boundaries between design, product, and engineering are being dissolved on purpose, while AI compresses the cycle further.

Timing prediction

Two comparable VP Design successions in the public record:

• Amplitude: Jenny Chang departed September 2024; Kim Lenox announced as VP Design December 2024. Three months.
• Datadog: Stephen Boak departed August 2025; external VP Design announced eight months later.

Sample size of two. The Amplitude case — external hire, enterprise SaaS, one-quarter turnaround — is the closer structural match.

The window: If Vanta follows a one-quarter cadence, the search moves from open pipeline to shortlist in three to five weeks, targeting a Q4 2026 start. Outreach that arrives after the shortlist forms competes against candidates who already have interview momentum and internal advocates.

What breaks this prediction: An internal candidate from the director bench already in process, which compresses the timeline. Or a retained firm that began sourcing before the posting went live, which means the pipeline is further along than five weeks suggests. No public evidence for either. I cannot rule them out.

The risk you name in the room

Vanta is shipping 100+ feature updates per quarter. Agents are expanding into write operations, sandbox execution, risk-register modifications. The engineering team is documenting new failure modes in public. The stated quality principle is "human in the loop, always."

With no VP Design and no publicly identified interim, the question of who holds quality across 90 planned agentic workflows falls to the director bench or to Epling himself. That is the specific risk the new hire absorbs on day one — and your strongest positioning angle. Name this risk, demonstrate you've solved a version of it, and you start the conversation where Epling needs it to start.

Decision-maker

Jeremy Epling, CPO. Three years in seat. In a First Round interview, he described wanting a VP Design who pushes back on craft. He told a story about shipping a "B class" product early in his career and receiving direct executive criticism. He runs weekly visual deep dives that move quickly to customer experience and prototypes.

He evaluates for: quality bar he respects, comfort operating inside a blended model without needing a separate fiefdom, ability to govern AI-accelerated output. He values designers who own front-end implementation and prototyping over specification.

Competitive landscape

Enterprise design director (high confidence). The default pipeline candidate — someone who's run 40+ designers at a growth-stage SaaS company. Strong on org management. Weak on AI-native credibility and the blended-model philosophy Epling has been building toward for three years.

AI-native design leader (moderate confidence). Someone from Anthropic, OpenAI, or a well-funded AI startup. Strong on AI fluency. Likely weak on enterprise quality-systems evidence and the org scale the posting demands. The 40-person team and director bench are real requirements.

Internal candidate (speculative). Kowitz, Sheehan, Larson, or Chang. No public signal. If one is in process, the external search may be validation rather than open competition.

Your differentiation: You map to both the AI-native and enterprise dimensions simultaneously. The Trust essay speaks directly to "human in the loop, always." TinyFish gives you practitioner depth on agent quality — traces, auditability, governance — that portfolio-only candidates don't have. Alibaba proves enterprise scale. Your vulnerability: you haven't run a 40-person design org, and the director bench management expectation is real. Lead with the quality-governance angle. Let the org-size question come to you rather than volunteering it.

Rubric

Narrative fit: Vanta's agents are writing to risk registers and executing in sandboxes while the quality-governance seat is empty. You've published the framework for where human oversight belongs in agentic workflows and you're building it in production at TinyFish. Alibaba covers the enterprise-scale proof the director bench needs above them.

Verdict: Act.

Outreach artifacts

1. First-contact message (LinkedIn DM to Epling)

Jeremy — your engineering team's recent post on sandbox failure modes (partial dataset inspection, authorization boundary crossings, unnecessary tool invocations) caught my attention because I'm working the design side of that exact problem at TinyFish, where I'm Head of Product building enterprise web agents in production.

The hardest design challenge underneath this isn't the interface on top of the agent. It's deciding where human oversight has to be designed into the workflow versus where it can be earned through demonstrated reliability. I published a framework for this — "Trust Is the New Interface," five handoffs where that decision gets made — and built the agentic design cases behind it at junochen.com.

Would 20 minutes be useful to compare notes on how you're thinking about quality governance across the 90-workflow target, particularly where the design function's role begins and ends in a blended org?

— Juno Chen, junochen.com

2. Cover letter draft

The hardest design problem at Vanta right now sits underneath the agent interfaces: deciding where human judgment is irreplaceable as agentic workflows expand from 20 to 90+, while your engineering team is documenting new failure modes from sandbox access and write operations in real time.

I've been working this problem from both the enterprise and AI-native sides. At Alibaba.com, I led design and research for the North America B2B platform ($50B+ GMV, +20% daily transactions, +2.2pt NPS) — enterprise-scale trust architecture where a buyer commits six months of inventory on the strength of a screen. That work taught me how to name structural gaps, build executive mandate, and redesign trust signals across multiple product surfaces simultaneously. I carried that into AI: at TinyFish (enterprise web agent platform, Series A), I'm Head of Product deploying agents in production daily — traces, auditability, governance at enterprise scale, three products shipped in three months. At Agentic Labs, I designed three enterprise agent interfaces live at junochen.com, each demonstrating the trust handoffs that determine whether a human can delegate or needs to stay in the loop. The framework behind this — "Trust Is the New Interface" — maps five handoffs and a trust ladder (Watch, Verify, Delegate) that tracks directly to Vanta's "human in the loop, always" principle.

I'm returning to design because I want to apply the technical depth I built at TinyFish to a high-stakes vertical domain where agent errors carry real consequences. Vanta's GRC problem is exactly that — and the quality-governance seat is open at the moment it matters most.

Portfolio and framework at junochen.com.

3. Resume emphasis guidance

• Lead with: Agentic Labs (all three cases) + Trust essay. Direct map to Vanta's agent-quality problem and Epling's blended-model philosophy.
• Second: Alibaba.com. Enterprise scale, trust architecture, cross-functional mandate-building. Proves the organizational altitude the posting requires.
• Third: Thermo Fisher mySupply. Regulated enterprise, exception-first design, 0→1 function build. GRC is a compliance domain — this case proves you've designed for environments where errors have regulatory consequences.
• TinyFish: Header: "Head of Product, TinyFish (enterprise web agent platform, Series A)." One bullet: "Shipped 3 products in 3 months; work daily with agent traces, auditability, and governance in enterprise deployments."
• Trim or move down: Equinox+ (consumer, less relevant). Red Cross only if the mission-critical angle comes up in conversation.
• Transition narrative: "Moved to product to build AI-natively from zero; returning to design to apply that technical depth to a high-stakes vertical domain."

Engagement intelligence

Ask in conversation: "How are you thinking about the VP Design's role in launch review as agent capabilities expand into write operations — is that a design-owned gate, or does it live across the blended org?" This surfaces whether Epling wants the VP Design to own quality governance or share it, which determines the actual scope of the role.

Don't say: Anything that frames design as needing protected territory. Epling's philosophy is explicitly about dissolving those boundaries. A candidate who argues for design's independence is arguing against his operating model.

What competing candidates will miss: Vanta's engineering team published specific quality degradation modes from sandbox support — partial dataset inspection, authorization boundary crossings, unnecessary tool invocations. Most candidates will pitch AI credentials generically. You can name the failure modes Vanta is actually experiencing and map your framework to each one.

Warning signs

Design influence ceiling scored 2. Design reports to the CPO, not the CEO. Epling values craft, but the blended model means design doesn't hold autonomous strategic authority. Your influence runs through him.

The stale careers page. Kawamoto still listed as VP Design five weeks after the posting. Could be web-team lag. Could signal the transition is being managed quietly while an internal candidate is evaluated.

The blended model in practice. "Designers submitting pull requests" can mean design is evaluated on engineering output metrics rather than design quality metrics. Ask how design quality is measured independently of shipping velocity.

Post-outreach signals

Vanta sits between growth-stage and enterprise. Apply growth-stage timing:

• Response within 3–5 business days = normal. Epling runs a large org and likely triages messages in batches.
• Silence past 7 business days = message didn't cut through, or pipeline is further along. One follow-up at day 8 referencing a new signal — a product launch, a post Epling makes, a new agent capability announcement.
• Recruiter reply instead of Epling = message landed but was routed to process. Worth pursuing; get to Epling in the screen.
• Two unanswered messages = redirect energy. Monitor for posting refresh or language change.

Warm path

No mutual connections surfaced. This is cold outreach. Epling's recent public appearances (First Round interview, Gupta episode) and the engineering team's sandbox post give you current-signal hooks that make cold feel informed. The sandbox/quality-governance angle references something his team published within the last month — it signals you're reading Vanta's engineering reality, not the job posting.