Principal Researcher, Facebook
2 papers at NeurIPS 2025
The first feed-forward reconstruction model for scene-level deformable 3DGS from a monocular video
Multimodal dataset and method to detect reading activity for applications towards AI glasses