We’re looking for someone who sees AI as a lever for fundamentally changing how product security operates — beyond scanning and ticket triage, toward systems that understand architectural context, reason about risk, and drive remediation. You’ll report to the Head of Security and operate with broad ownership and real autonomy.
You will
Own product security across the full lifecycle — secure design, threat modeling, implementation, release, operation, and decommissioning — for everything we ship.
Build application security into engineering workflows: secure defaults in frameworks and libraries, code and secret scanning that produces meaningful signal in CI/CD, and guardrails developers adopt because they’re useful.
Own vulnerability management end to end — findings from code scanning, dependency and secret scanning, pentests, and our bug bounty program — with clear severity calibration, SLAs, and fixes driven to closure.
Own software supply chain security — dependencies, build and release integrity, artifact provenance, and third-party code and services inside our trust boundary.
Fix vulnerability classes rather than instances: trace findings to root cause and eliminate the pattern at the framework or platform level.
- Build AI-driven tooling that reviews architecture, triages findings, and drives remediation with minimal human intervention — and know where human judgment still belongs.
- You may be a good fit if you:
- Have 7+ years across product security, application security, or security engineering — with hands-on ownership of secure design, threat modeling, and vulnerability remediation for a production product.
Write real software. You read the codebase fluently, ship the fix yourself, and build internal tooling where commercial scanners fall short — and know how to tune the ones you keep so engineers trust the signal.
Have owned software supply chain and SDLC security — dependencies, build integrity, CI/CD controls, and third-party risk inside the trust boundary.
Have run vulnerability management or a bug bounty program — triage, honest severity calls, researcher communication, and getting fixes shipped on SLA.
Think in systems and lifecycles, not findings. You eliminate vulnerability classes and reason about what happens to data and access when something is retired.
Have shipped security controls that engineers actually adopted, and can explain why the ones that failed failed.
Have already built something with AI that changed how you work — agentic automation, not autocomplete — and have opinions about where human judgment still belongs.