I work on gemini at Google DeepMind, where I lead the frontier evals team.
I’m interested in the future of AI R&D automation and the technical challenges of evaluation, oversight and reward hacking in large scale RL runs.
Previously, I was an assistant professor at the University of Cambridge. Before that, I was a researcher in the Visual Geometry Group and a research fellow at Balliol College.
I’m based in King’s Cross, London’s techno-optimist postcode.

