Tom McGrath

Tom McGrath is chief scientist and co-founder of Goodfire, an AI interpretability startup. He previously served as a research scientist at Google DeepMind from 2019 to 2023, where he co-founded the interpretability team and worked on understanding the internal mechanisms of models including LLMs and AlphaZero. He holds a PhD in mathematics and statistics from Imperial College London and is known for research on topics such as the Hydra Effect, copy suppression, and the science of training data.