- Where
- Remote: Remote, country not statedRemote, country not stated
- Level
- Early–Mid career, 1–9 yrsEarly–Mid1–9 yrs
- Pay
- Not listed
- Apply by
- Rollingposted 10 Sep
- Hours and type
- Full-time
- Kind of work
- Technical & Engineering
- Added to CFI
- 10 September 2026
- Network
About the role
Attacks the cyber safeguards of frontier AI models, designing new attack strategies and documenting what breaks. Findings go to AI labs to shape system cards and deployment decisions.
Summary of the employer's page written by AI. Check the full listing before you apply.
Requires
Successful jailbreaks of frontier AI models; security mindset; sign a confidentiality agreement
This role's deadline has passed, so it is no longer on the board. See open roles.