Agents write the code. You decide what gets built, how it's verified, and when it ships.
About the role
As Principal AI Engineer, you are the most senior technical engineer in our embedded client teams. The role is fully hands-on, with no people management (market equivalents: Staff / Principal Engineer). You'll work embedded with the client and report to the Tech Director.
On each engagement, you'll act as Forward Deployed Engineer, Agentic Tech Lead, or both. When the team needs it, you'll also take on other technical roles.
What you'll do
Own a workstream end to end (FDE)
- Lead a workstream through discovery, delivery, and integration into the client's systems
- Hand over operations to the client's team at the end of the project, so they can run it independently
Architect and verify agents (Agentic Tech Lead)
- Design agents, defining each one's purpose, scope, and evals
- Verify at scale: agents write the code, while judgment and sign-off stay with you
- Train the client's team in agentic engineering patterns
Stay flexible
- Take on other technical roles as the team's allocation requires
What this role is not
This is not a people-management role, and you won't lead the team; that's the Tech Director's job. The role is hands-on by definition: you architect, build, and verify.
How we work
- Agentic harness: Claude Code and Cursor. You build and operate the harness that runs agents in production.
- Agent architecture: MCP, subagents, and verification loops, with a defined purpose, scope, and evals for every agent.
- Evals at scale: blocking evals in CI, a sandbox per agent, and execution telemetry (tool calls, cost, latency).
- Spec-Driven Development: an executable spec governs every change, and every agent PR gets human sign-off.
- Experimentation & FinOps: A/B testing in production, guardrails, prompt caching, model routing, and tracking token cost per delivery.
What success looks like
- Your workstream is delivered and handed over, and the client runs it with no dependency on us
- Agents are running in production under blocking evals
- Human sign-off holds up at volume
- The client's team is operating on agentic patterns
What we're looking for
- 4-6+ years of senior engineering experience shipping production software
- Experience with technically complex systems or products: multiple integrations, scale, and real-world production constraints
- You've architected and operated agents in production (purpose, scope, evals), not just used them
- You review agent-generated PRs rigorously and stand behind your sign-off
Nice to have
- Experience with experimentation platforms, A/B testing, or GenAI products
- Experience delivering while embedded with clients
- A background as a Staff Engineer with founder or deep-tech experience
- Experience as a Tech Lead or Staff Engineer with production sign-off
You’ll thrive here if you...
- Show ownership beyond your job title, through side projects, founding experience, or initiatives nobody asked for
- Challenge scope and propose better paths instead of executing blindly
- Are in a growth phase of your career and value challenge and upside over stability
- Unblock yourself without waiting for direction
- Have a track record of taking on responsibility above your title