Trigger
HackerOne shipped three products in ten weeks: H1 Platform (June 2), H1 Remediation (July 29), and a Remediation Dashboard (August 2026). Each one creates a new human-authority surface for autonomous security agents. CPO Nidhi Aggarwal is building these faster than the design team can architect them. The Director, AI Product Design posting reported directly to Aggarwal, was confirmed accessible on August 1, and is now unlisted from both Ashby and LinkedIn. Confidence: high on product velocity, moderate on search status. Unlisted could mean filled, paused, or moved nonpublic. The design need didn't disappear with the posting.
Window
The posting's disappearance narrows the application window but widens the outreach window. If the first search failed, a message that names the design problem Aggarwal is publicly wrestling with lands better than a job application ever would. If it filled, you've lost five minutes.
Aggarwal posted about the discovery-to-remediation gap roughly a month ago and is still building into it. The product releases give you a durable reason to reach out regardless of posting status.
Act on outreach this week.
What this trigger means for the design org
HackerOne's leadership page lists no Head of Design, VP Design, or equivalent. The previous Head of Product Design left for Invisible Technologies. The one before that went to WeTravel. The visible design bench is a Staff AI Product Designer and a Lead Product Designer. The Director posting described leading and scaling the design function, reporting to the CPO directly.
This is a function-building mandate. Aggarwal needs someone who can stand up design leadership for a product surface that is multiplying:
- Agentic Validation — accept/modify/reject for vulnerability disposition
- Asset Intelligence — approve/adjust/decline for scope changes
- MCP server — enforces read-only access
- Remediation — auto-generates fix plans
Four authority levels across four actions, and nobody currently owns the coherence across them.
Your angle
Lead with Carrier IQ and the Trust essay. HackerOne's accept/modify/reject disposition model is the most specific commitment-point interaction in your current target set. Your published framework names what it is: a Decision Gate where the human retains authority over an agent's recommendation before consequences flow. Carrier IQ demonstrates this in a regulated domain — InsurTech quote automation where the agent completes the form and the human owns the coverage decision. The Trust essay's five handoffs and trust ladder (Watch, Verify, Delegate) map to HackerOne's graduated authority model across its product surface.
Secondary: Thermo Fisher and Red Cross. These counter the domain gap. You have no cybersecurity employment. You do have regulated platform design where a wrong decision has immediate downstream cost — pharma supply chain exceptions that cost 3–5× more at the delivery gate, disaster relief disbursement where the wrong determination delays aid. The design problem at the commitment point transfers across these domains.
The argument rests on conditions that apply regardless of industry: does the human at the gate have visible state and provenance, sufficient time and training, recognized authority, the ability to stop or reverse the system, and no penalty for exercising that authority? HackerOne's documentation shows they've built the gate. They need someone who can design the conditions around it.
Who to contact
Nidhi Aggarwal, CPO. LinkedIn.
Public fingerprint: Active on LinkedIn in the last 90 days. Her strongest recent framing (~2 months ago): validation must lead into remediation without losing context between security and engineering, and repeated vulnerability classes should be treated as a design problem rather than a queue of tickets. A ~1-month-old post cited HackerOne data showing submissions up 76% but resolved volume down 46%, with the unresolved backlog at 21× monthly resolution capacity. She also participated in a June 25 session on continuous threat exposure management and making security decisions that withstand board scrutiny.
"Collapse the handoffs" and "design problem, not a queue of tickets."
Warm path: No shared VC connections, ADPList mentors, or BCG DV / Alibaba network overlaps surfaced. This is cold. Her recent LinkedIn activity gives you a specific hook.
Channel: LinkedIn cold DM. Her posting frequency suggests she reads the platform.
Secondary: Joshua Dunne, Staff AI Product Designer. LinkedIn. He's likely closest to the authority-surface design work right now. Don't reach out before Aggarwal — she's the decision-maker and the posting reported to her. If Aggarwal engages, Dunne becomes a useful intelligence source: what the current design process looks like for Agentic Validation, how decisions about the accept/modify/reject model actually get made, where the team feels the gap. Reference him in conversation with Aggarwal if it helps demonstrate your read on the org.
Outreach messages
Cold message:
Your post about treating repeated vulnerability classes as a design problem — not a ticket queue — is the sharpest framing I've seen of what happens when validation and remediation lose context between handoffs.
I've been working on this exact seam. I published a framework for human-authority preservation in agentic systems ("Trust Is the New Interface," junochen.com) and built the design for an InsurTech agent that automates quote completion while preserving human authority at the coverage decision — the same accept/modify/reject commitment point your Agentic Validation documents.
I'm currently Head of Product at TinyFish (enterprise web agents, Series A), working daily with agent traces, auditability, and governance in production. I'd welcome 20 minutes on how you're thinking about design leadership for the graduated authority model across H1 Platform — the different control surfaces for disposition, scope, and remediation feel like they need a unified design architecture.
Warm message (if a connection surfaces):
[Name] suggested I reach out. Nidhi's framing of repeated vulnerability classes as a design problem — and the need to collapse handoffs between validation and remediation — maps directly to work I've published and built. I designed the decision gate for an InsurTech agent that automates quote completion while preserving human authority at the coverage commitment — the same accept/modify/reject structure documented in Agentic Validation. The framework behind it is at junochen.com. I'd welcome 20 minutes on how she's thinking about unifying the authority model across H1 Platform's different control surfaces.
What not to say:
- Do not reference the Director posting. It's unlisted. Mentioning it signals you're tracking the job board, not the design problem.
- Do not lead with Alibaba or enterprise scale. Aggarwal's language is about collapsing handoffs and workflow architecture. Org size is not the test here.
Cover letter draft
Use only if the posting reappears or Aggarwal directs you to apply formally.
HackerOne's product surface now documents at least four distinct authority levels — accept/modify/reject for vulnerability disposition, approve/adjust/decline for scope, read-only for MCP, automated generation for remediation plans. Each answers the same design question differently: where does the human retain authority, and what do they need to exercise it well? That graduated model needs a unified design architecture.
I designed the decision gate for Carrier IQ, an InsurTech agent that automates quote completion while preserving human authority at the coverage commitment — the same structural problem as Agentic Validation's disposition model. The underlying framework is published: "Trust Is the New Interface" (junochen.com) maps five handoffs and a trust ladder that describes what HackerOne is already building across its product. Before that, I built regulated platforms where wrong decisions have immediate downstream cost — Thermo Fisher's pharma supply chain ($20M+ margin recovered, 6 partners, exception-first design) and American Red Cross disaster relief ($847K disbursed, 6 systems consolidated to 1, national deployment).
I'm currently Head of Product at TinyFish, an enterprise web agent platform (Series A), where I shipped 3 products in 3 months and work daily with agent traces, auditability, and governance in production. I moved to product to build AI-natively from zero; I'm returning to design because the hardest problems in agentic systems are design problems — exactly the framing Nidhi has been articulating publicly.
I'd welcome the chance to discuss how I'd approach design leadership for H1 Platform's authority surfaces. Portfolio and framework: junochen.com.
Resume emphasis guidance
- Lead with Carrier IQ. The InsurTech quote automation case maps directly to HackerOne's accept/modify/reject model. Place it first under a clear Agentic AI section.
- Trust essay second. List as a published work with the five-handoffs framework named. The graduated authority model across HackerOne's product is the trust ladder in production.
- Thermo Fisher third. Domain-gap counter — regulated platform, exception-first design, $20M+ margin, 0-to-1 function build. Proves you can design for consequential decisions without cybersecurity employment.
- Red Cross fourth. Same logic — high-stakes, multi-system consolidation, compliance constraints.
- TinyFish in the header as bridge: "Head of Product, TinyFish (enterprise web agent platform, Series A)." One line about shipping 3 products in 3 months and daily work with agent traces and governance.
- Trim or move down: Alibaba, Equinox+, Allē. Enterprise scale, consumer speed, and dual-surface work are all secondary to the authority-preservation story here.
Post-outreach signals
HackerOne is late-stage but the design org is startup-sized. Treat this like growth-stage:
- Response within 3–5 business days = normal.
- Silence past 7 business days = one follow-up at day 8 referencing a new signal (a product release, a post from Aggarwal). No reply after follow-up, move to Watch.
- Aggarwal replies engaging the design problem you named = respond within 24 hours and deepen the thread on the graduated authority model.
- Recruiter replies with a generic screen request = the message landed but didn't cut through. Still worth pursuing. Get to Aggarwal in the screen.
- A reply requesting work samples or a portfolio walkthrough = high-intent signal. Before the conversation, prepare the junochen.com walkthrough framed around the graduated authority model — Carrier IQ as the lead case, Trust essay as the architecture, the four HackerOne authority surfaces as the application layer.
- Calendar availability sent in the first reply = move HackerOne to the top of your daily queue. This search is moving.
Competitive candidate landscape
Cybersecurity design leader archetype (moderate confidence). The pool is thinner than the domain-gap objection implies. Across CrowdStrike, Palo Alto Networks, SentinelOne, Snyk, Wiz, and Fortinet, only CrowdStrike has a publicly named Director+ design leader. CrowdStrike is actively posting for a Director-level design operations role, and both Palo Alto Networks and SentinelOne recently ran Director-level design searches — which suggests the senior cybersecurity design bench is shallow. Aggarwal may not have a cybersecurity-native design director available in the pipeline.
AI-native design leader archetype (moderate confidence). Candidates from Anthropic or OpenAI will carry stronger AI brand recognition but almost certainly lack published supervision logic mapped to a specific commitment-point interaction. They'll talk about responsible AI in general terms. You can talk about accept/modify/reject in specific ones.
Enterprise design director archetype (speculative). Alibaba/Salesforce pedigree candidates bring org-scale proof but will struggle to demonstrate hands-on AI-agent design or consequence literacy in adversarial domains.
Your vulnerability is the domain gap. You've never worked in cybersecurity. The counter-move is to reframe the test: the design problem at the commitment point transfers across regulated domains. Visible state, sufficient time, authority, reversibility, no penalty for exercising judgment. You've designed this in pharma, disaster relief, and InsurTech. The adversarial context adds urgency to the gate design but doesn't change the gate's structure.
The non-obvious thing
HackerOne's product documentation reveals a graduated authority model that the company hasn't publicly named as a unified design pattern. Accept/modify/reject for disposition. Approve/adjust/decline for scope. Read-only for external agents. Auto-generated plans for remediation. A candidate who walks in and names this pattern — "you're already building a trust ladder across your product surface, and I wrote the framework that describes it" — demonstrates a read on the product that no other candidate is likely to have done. The documentation is all public. Nobody has synthesized it.
Warning signs
The posting is unlisted. You're reaching out without a confirmed open role. Frame your message around the design problem, not the job.
Design has not had executive representation at HackerOne. Two previous Heads of Product Design left. The leadership page lists no design executive. This could mean Aggarwal genuinely wants to elevate design, or it could mean the Director title is the ceiling. First-question diagnostic if you get a conversation: ask where design decisions about the authority model currently get made. If the answer is product management or engineering, you're being hired to fight for a seat.
No cybersecurity employment in your background. Aggarwal may weight domain expertise more heavily than the public signal suggests. If the first search failed because they couldn't find a cybersecurity-native design leader at this level, you're well-positioned. If it failed because they decided they need deeper domain knowledge, you're not. You cannot know which from public signal alone. Reach out anyway.
- Aggarwal's remediation gap data: Her ~1-month-old LinkedIn post cited an unresolved backlog at 21x monthly resolution capacity — if she posts updated figures or a follow-up, that's your day-8 follow-up hook.
- H1 Remediation's missing gate: The July 29 launch material describes auto-generated fix plans routed into engineering tools but does not document who approves the generated code change before merge — a gap worth raising in conversation as evidence of your read on their authority model.
- CrowdStrike's design operations search: Their active Director-level posting suggests CrowdStrike is building out its own design leadership bench, which further thins the cybersecurity design-leader pool available to HackerOne.
- Asset Intelligence scoping controls: HackerOne's June changelog documents approve/adjust/decline for AI-generated scope recommendations — a second commitment-point surface that reinforces the graduated authority pattern and gives you additional product specificity in conversation.