AI Safety & Alignment jobs

AI safety and alignment has grown from a niche into a working discipline: dangerous-capability evals, red-teaming, interpretability research, and the policy-adjacent roles that translate all of it for regulators. The titles to watch for are Alignment Researcher, Safety Engineer, and Model Policy Lead, with trust-and-safety hybrids appearing at product companies.

The surprise for many applicants is how much of the work is engineering — safety teams build eval harnesses and monitoring infrastructure, not just position papers, so RLHF literacy and solid ML plumbing both matter. It's a small field concentrated at a handful of labs, which keeps openings scarce and competition real. Remote policies vary lab by lab; pay is strong and rarely printed.