I am a researcher at Google DeepMind working on AI alignment. My research aims to develop safe, interpretable, and trustworthy AI systems that learn from human feedback.

I received my PhD from ETH Zurich, where I was part of the Learning & Adaptive Systems Group supervised by Prof. Andreas Krause and Dr. Katja Hofmann. My dissertation focused on developing “Algorithmic Foundations for Safe and Efficient Reinforcement Learning from Human Feedback”.

Research

My research focuses on what I believe are key ingredients for building safe and trustworthy AI systems: (1) learning from human feedback safely and sample efficiently; (2) providing robust supervision for very capable AI systems (e.g., via scalable oversight); (3) evaluating AI systems for dangerous capabilities; and, (4) monitoring and red-teaming of AI systems. My research interests tend to shift between these areas. To learn more about my current work, see my recent publications.

News

• July 2024: I gave a talk at the ICML Models of Human Feedback for AI Alignment Workshop.
• July 2024: New paper on arXiv: “On scalable oversight with weak LLMs judging strong LLMs”.
• March 2024: New paper on arXiv: “Evaluating Frontier Models for Dangerous Capabilities”.
• February 2024: “Vision-Language Models are Zero-Shot Reward Models for Reinforcement Learning” was accepted to ICLR 2024.
• January 2024: “Learning Safety Constraints from Demonstrations with Unknown Rewards” was accepted to AISTATS 2024.
• December 2023: I’ll be at NeurIPS 2023 in New Orleans presenting Tracr at the main conference and our work on VLM-RMs at the FMDM Workshop.
• October 2023: New preprint on arXiv: “Vision-Language Models are Zero-Shot Reward Models for Reinforcement Learning”.
• September 2023: Our paper “Tracr: Compiled Transformers as a Laboratory for Interpretability” was accepted to NeurIPS 2023 as a Spotlight.