I'm an AI researcher in the field of mechanistic interpretability: developing mathematical methods for opening up the black boxes of state-of-the-art AI systems, in order to understand when and why they do things that we don't want.

Previously, I've also worked on optimizing performance in cloud-scale distributed systems and researching network traffic optimization -- but I enjoy tackling any interesting problem I can get my hands on. Other interests include mathematics, jazz piano, and linguistics.