Research has been published at CVPR, ICCV, ECCV, and ICLR.

I am currently a researcher at AMI Labs, where I aim to enable human-like visual understanding and learning in machines. Prior to AMI, I did research into self-supervised learning, video representations, and multimodal modeling at FAIR with the JEPA team. I was a core contributor to V-JEPA 2 and MetaMorph, and a lead for WebSSL and Beyond Language Modeling.

As of March 2026, my open-sourced models from FAIR have over 1M total downloads and 500 citations. Prior to FAIR, I was at Amazon Prime Video, where I worked primarily on video/multimodal representation learning with a long-term vision of cracking the "long"-video understanding problem.

My notable production work includes: 1) Prime Video automated advertisement insertion at scene boundaries, 2) semantic-knowledge graph for efficient zero-shot metadata classification at catalog scale, 3) visual-search for carousel recommendations, and 4) content moderation and compliance models.

I graduated magna cum laude from Princeton University with a B.S.E in computer science in 2019. My life's ambition is to empower people to do their best work and strive to be better versions of themselves.