Description
The AI Security Institute is hiring Research Engineers and Research Scientists for its Alignment Red Team to research and evaluate misalignment risks in frontier AI systems, including loss-of-control risks such as research sabotage and reward-seeking. The role involves developing alignment evaluation methods, running pre-deployment tests, investigating misaligned behaviour, contributing to research publications, and building software and tooling. Candidates need substantial AI safety or alignment research experience, strong software engineering and machine learning experience, and at least one year of professional Python experience. The role offers hybrid working, a modern central London office, and annual salaries ranging from £65,000 to £145,000 depending on level.
