Lee Sharkey

Lee Sharkey is an experienced researcher in mechanistic interpretability, focusing on understanding the learned behaviors of neural networks to ensure their safe deployment. They worked as a Teaching Assistant at ETH Zürich and as an Independent AI safety researcher, where they authored a review on deep reinforcement learning agents. Lee has held various roles, including Research Engineer at Conjecture, where they published multiple research outputs, and is currently a Principal Investigator at Goodfire. They hold advanced degrees in Data Analytics, Neural Systems and Computation, and have extensive experience with deep learning frameworks.