Sham Kakade

Co-director of the Kempner Institute
Gordon McKay Professor of Computer Science and Statistics

Preferred Pronouns:


Contact Information


Sham Kakade, who joined the Harvard University faculty in spring 2022, works on the mathematical foundations of machine learning and AI. He focuses on the design of provably efficient and practical algorithms that are relevant for a broad range of paradigms. Kakade’s thesis helped lay the statistical foundations of reinforcement learning; he helped to pioneer spectral and tensor methods for latent structure discovery and he has been a pioneer in the analysis of stochastic gradient methods in nonconvex optimization problems, such as those that arise in training deep neural networks. Kakade also developed the mathematical foundations for the widely used linear bandit models and the Gaussian process bandit models, work for which he received the Test of Time Award at the 2020 International Conference on Machine Learning. He earned his PhD at the Gatsby Computational Neuroscience Unit at the University College London and came to Harvard from the University of Washington, where he was a professor in computer science and statistics. He has also been a principal research scientist at Microsoft Research in New England and New York City.

Research Focus

Kakade’s focus is on developing efficient and practical algorithms for foundation models and real-world AI applications. His current research interests include: (1) optimization with an emphasis on real engineering challenges, (2) exploring the foundational science and mathematical underpinnings of deep learning, and (3) advancing the usefulness of LLMs and generative AI.