I am a Research Scientist at Google DeepMind in San Francisco, where I work on Gemini post-training for reasoning and coding. My current research focuses on reinforcement learning for large language models, test-time compute, and train-time search.

Previously, I was an associate professor (status-only) in the Department of Electrical & Computer Engineering at the University of Toronto, a faculty member at the Vector Institute for Artificial Intelligence, and a Canada CIFAR AI Chair from 2018 to 2024. My research there spanned reasoning, alignment, and inference in LLMs; deep generative models (particularly diffusion models); variational inference and Monte Carlo methods; computational optimal transport; optimization for neural networks; and the intersection of machine learning with information theory.

I completed my PhD at the University of Toronto with Brendan Frey in 2018, as part of the Machine Learning Group. During my PhD, I interned at Google DeepMind (2016) and Google Brain (2015). I received my Master's degree from the University of Toronto in 2012 and my Bachelor's degree from Amirkabir University of Technology, Iran, in 2010.

Selected Publications