Thomas Wyndham Bush

About

Thomas Bush is a computational researcher working at the intersection of machine learning, neuroscience, and biology. He studied Data Science at Tilburg University and Cognitive Science at the University of Trento.

Bush conducted research at the Iurilli Lab (Italian Institute of Technology), where he contributed to computational pipelines for parsing freely moving animal behavior. He later worked as a private contractor for the European Union Drugs Agency, supporting the development and evaluation of AI tools for research workflows. In 2025, he received an Armenise Harvard Summer Fellowship to join the Sabatini Lab at Harvard Medical School.

Research Focus

Thomas Bush’s current research focuses on building computational tools to study the functional effects of protein mutations. His work develops machine learning pipelines that predict sequence-to-function relationships in proteins, with the goal of understanding how changes in amino acid sequences shape stability, structure, and measurable functional readouts.

A key emphasis of this research is methodological: benchmarking modern structure- and sequence-based models, designing robust evaluation protocols, and developing interpretable representations that reveal which mutations drive functional changes and why. Ultimately, this line of work aims to contribute scalable computational methods for protein engineering and for the mechanistic interpretation of mutational landscapes.

More broadly, his main machine learning interests are in representation learning and interpretability for biological data, with particular attention to robust evaluation and reproducible research software. He enjoys building open-source tooling and Linux-centered workflows that enable scalable experimentation.