I'm an Assistant Professor of Computer Science at Princeton University. My research builds AI systems that perceive and understand the world, studies how models work, and explores how to use them to accelerate scientific research.

I value simplicity, and look for clean models, methods, and ideas that make complex systems easier to build and understand. My work has introduced simple, widely used designs in deep networks and efficient models (e.g., DenseNet, ConvNeXt, Network Slimming, Rethinking Pruning, Wanda), and more recently, simple designs and analyses that help explain how multimodal and large language models work (e.g., DyT, MetaMorph, Massive Activations, dataset bias). My work has been recognized by 110,000+ citations and a CVPR Best Paper Award.

Before joining Princeton, I was a Senior Research Scientist at Facebook AI Research in New York. I received my Ph.D. from UC Berkeley, where I worked with Trevor Darrell, and my B.S. in Computer Science from Tsinghua University (Yao Class).