David Duvenaud

David Duvenaud is an associate professor in the Department of Computer Science at the University of Toronto. His research focuses on machine learning, probabilistic models, and neural networks. His recent work includes AGI governance, evaluation, and mitigating catastrophic risks from advanced AI systems, following a sabbatical with Anthropic. Previously, he contributed to deep probabilistic models, including neural ODEs, generative chemistry design, and hyperparameter optimization. He is a founding member of the Vector Institute and holds the Schwartz Reisman Chair in Technology and Society.

David Duvenaud is an Associate Professor at the University of Toronto, specializing in areas such as LLM Evals, Differential Equations, and Approximate Inference. His research focuses on developing novel methods for modeling complex systems using neural networks and probabilistic approaches. He has contributed significantly to the field of deep learning through works like Neural Ordinary Differential Equations and Latent ODEs for irregularly-sampled time series. His publications also explore topics such as variational autoencoders, reversible generative models, and energy-based models. Duvenaud's work bridges theoretical and applied aspects of machine learning, with applications in chemistry, molecular design, and structured representations.