Summary
Takuya Ito is a research scientist at IBM Research with 15 years of experience at the intersection of artificial intelligence and computational neuroscience, specializing in transformers, reasoning, interpretability, compositionality, and biophysical time-series modeling. He combines rigorous theoretical training (PhD in Neuroscience) with hands-on ML research, having held postdoctoral positions at Yale and IBM and contributed to IBM’s Mathematics of Computation group. His work bridges neural data analysis and modern NLP architectures, emphasizing mechanistic understanding of model behavior rather than purely empirical gains. Takuya’s background in mathematics and computer science underpins a quantitative approach to problems in representation and dynamics, informed by publications spanning both AI and neuroscience. Based in New York, he maintains an active research presence with an accessible publication catalog and website that highlight cross-disciplinary collaborations and tooling. Colleagues value him for translating biophysical insights into interpretable ML methods that advance both neuroscience hypotheses and practical model diagnostics.
15 years of coding experience
8 years of employment as a software developer
B.A. Mathematics; Minor in Computer Science, B.A. Mathematics; Minor in Computer Science at Washington University in St. Louis
Doctor of Philosophy (Ph.D.) Neuroscience, Doctor of Philosophy (Ph.D.) Neuroscience at Rutgers University