October 2024: Core contributor to MovieGen, a state-of-the-art media generation system outperforming SORA and Emu Video, covered by NY Times, FT, Forbes, etc.
July 2024: Contributed to Llama 3.1 and state-of-the-art video recognition capabilities
June 2024: LaViLa and Ego4D received EgoVis 2022-23 Distinguished Paper Awards
November 2023: Announced Emu Video, a state-of-the-art video generation model, widely covered by TechCrunch, The Verge, etc.
May 2023: Co-developed ImageBind, a multimodal embedding model
June 2022: Presented OmniVore at CVPR 2022—a single model achieving SOTA across images, videos, and single-view 3D
October 2021: Launched Ego4D, the largest egocentric video dataset to date
Published multiple papers at CVPR, ICML, and arXiv on video generation, multimodal learning, and 3D representation
Research Experience
Research Scientist at Meta, New York, 2019–Present
Research Scientist Intern at DeepMind, London, Summer 2018
Research Scientist Intern at Facebook, Menlo Park, Summer 2017
Research Scientist Intern at Adobe, San Francisco, Summer 2016
Software Engineering Intern at Facebook, Menlo Park, Summer 2013