Research Associate, Detecting Dangerous Capabilities in Frontier ModelsNew
Imperial College London
Job AI safety & policy
- Where
- Work mode not stated: London · UKLondon
- Level
- Mid career, 5–9 yrsMid5–9 yrs
- Pay
- £51k–£59k£50,733 – £59,484
- Apply by
- 22 Novposted 23 Octin 17 days
- Hours and type
- Full-time
- Kind of work
- Research
- Added to CFI
- 28 October 2026Posted by employer 23 October 2026
- Network
About the role
Designs and develops algorithms to detect dangerous capabilities in frontier AI models that are not verbalized in their chains of thought. Publishes research, implements ideas and mentors PhD and MSc students in Imperial's Safety & Alignment lab.
Summary of the employer's page written by AI. Check the full listing before you apply.
Requires
PhD (or soon to obtain) in CS, maths, physics or similar; publications; PyTorch experience
This role's deadline has passed, so it is no longer on the board. See open roles.