I’m a PhD candidate, advised by Tim Rocktäschel and Edward Grefenstette, at the UCL DARK lab. I also work as a member of the technical staff at Anthropic, where I focus on building safe superintelligence. I’m interested in studying techniques to ensure Scalable Oversight. I consider these problems primarily in the setting of multi-agent learning. Our recent work on debate receieved a Best Paper Award 🏆 at ICML 2024.

My research is shaped by my AI safety concerns. I’ve developed machine-learning systems at Tractable and Spherical Defence Labs. These are active systems interacting and learning in the real world. I’ve also served as a grant maker at the Cooperative AI Foundation to mitigate the risks of multi-agent systems. For a more in-depth account, see my resume.

I’m more than just my research, feel free to explore my blog and see what excites me.