Research Engineer, AI Safety & Evaluations

1. Do you have strong hands-on software engineering experience and a serious interest in AI alignment, AI safety or security?
2. Have you built, evaluated or directed LLM agents while they perform complex technical tasks?
3. Can you independently own technical projects end to end, from initial idea through implementation, testing and refinement?
4. Do you have experience with Python, agent tooling, evals, RL environments or similar technical infrastructure?

If you answer yes to all of the above, then this confidential role could be the one for you.

This is a rare opportunity to join a small, fast-growing AI company working on technical alignment and evaluation problems for frontier models. The successful candidate will independently build environments and evaluation tasks that help identify misaligned or unreliable model behaviour, while improving the software used to isolate, measure and refine those behaviours.

The role is based in San Francisco. Salary on offer is $200,000 to $400,000