AI Researcher, AI Safety & Security
TinycodeRemote
Global catastrophic risksAI safetyResearchSoftware engineering
Why this is on the board
A good testing method isn't something you use once. If you find a way to measure how an AI system goes off the rails or can be misused, the labs building the most powerful models can rerun that test on every new model, as can other researchers in AI safety. Whether that makes systems safer does depend on what those labs and companies do with the results.
This board is in beta and this is our judgement. If you think we have it wrong, tell us. Give feedback · How we reason about it
About the role
Lead research on AI system behaviour, vulnerabilities, and defences through custom evaluations and autonomous red-teaming for enterprises and frontier labs. - Design and build end-to-end AI evaluations for labs and enterprises, from threat modelling to scoring and reporting. - Research AI failure modes, capability elicitation, and mitigations across models, RAG pipelines, agents, and multi-agent systems. - Advance autonomous red-teaming techniques and develop methods to assess evaluation covera…
Not the right role? Get new ones every month
On the first Tuesday of every month: new jobs and programmes from this board, plus events in the next six weeks and the best of the EA Forum.
Want to understand this first?
Start with our free course: about 100 minutes in six short units, at your own pace, with no sign-up. To go further, apply for the EA Fellowship: five weeks of reading and discussion in a small group. Also free, but selective.