Assran et al.

Mahmoud Assran is a research scientist at Meta AI's Fundamental AI Research (FAIR) lab in Montreal, Canada. He specializes in self-supervised learning and world modeling, with a focus on Joint Embedding Predictive Architectures (JEPA) that learn abstract representations in latent space rather than generating pixels. He is best known as lead author of V-JEPA 2 (2025), a large-scale video world model trained on over one million hours of internet video that achieves state-of-the-art results in motion understanding, action anticipation, and zero-shot robot control. He holds a PhD from McGill University and Mila and also co-authored I-JEPA, presented at CVPR 2023.