Research Engineer / Research Scientist, Model Transparency

A role for technical AI safety researchers and engineers to join the UK AI Security Institute's Model Transparency team, working to ensure that evaluations and oversight of frontier AI systems remain reliable as models grow more capable.

Mission and access – work within the UK government with direct influence on frontier AI governance, collaborating with Anthropic, OpenAI, DeepMind, METR, and senior government officials including the Prime Minister's AI Advisor
Research focus – investigate evaluation awareness, unfaithful chain-of-thought reasoning, and model transparency risks; develop white-box and black-box methods to detect and mitigate oversight failures
Compensation and benefits – salary range of £65,000–£145,000 (base + technical allowance), 28.97% employer pension contribution, pre-release access to frontier models, and funding for learning, conferences, and external collaborations
Flexible pathways – open to both Research Scientists (experiment design, analysis, publications) and Research Engineers (tooling, infrastructure, scaling workflows) across a range of experience levels

Applications close Sunday 25 May 2026. Apply directly at job-boards.eu.greenhouse.io.

This text was generated by AI. If you notice any inconsistencies, please let us know using this form

View opportunity

Related opportunities

Senior Software Engineer, Chem-Bio

AI Security Institute (AISI)London, United Kingdom

London, United Kingdom

3 weeks ago

Senior Software Engineer, Chem-Bio

AI Security Institute (AISI)London, United Kingdom

London, United Kingdom

3 weeks ago