JOB-2058

An early-stage company building autonomous AI agents for engineering teams in regulated industries and defense is looking for a founding security engineer to design and ship the security foundation that makes those agents safe to deploy.

Highlights of the job

  • Own the threat model and security roadmap, mapping trust boundaries and data flows and defining testable acceptance criteria
  • Build the policy and execution layers that sandbox generated code, privileged tools and extensions, with scoped permissions and approvals
  • Implement identity, short-lived credential handling, tenant isolation and controls over network egress, storage and retention
  • Create auditable action records and approval trails, and make sensitive operations fail safely
  • Harden desktop packaging, updates and customer-hosted or restricted-network deployments
  • Write adversarial tests, investigate vulnerabilities and support enterprise security reviews

What you bring

  • 4–10 years as a production software engineer, with substantial security controls you personally shipped
  • Strong CS fundamentals: algorithms, concurrency, operating systems and systems design
  • Hands-on experience with authorization, secrets, process isolation, permissions or secure execution, ideally including desktop application security
  • Clear thinking about AI-specific threats such as prompt injection, untrusted model output and hostile tool inputs
  • Bonus: Python, TypeScript or Rust; native desktop sandboxing and signed releases; agent runtimes or tool protocols; CAD/PLM integrations

On-site in San Francisco, five days a week, reporting to the CTO. Visa sponsorship available.