1. Seamless Interaction: Dyadic Audiovisual Motion Modeling and Large-Scale Dataset
2. InteractAvatar: Modeling Hand-Face Interaction in Photorealistic Avatars with Deformable Gaussians
3. Visualizing Nanoparticle Surface Dynamics and Instabilities Enabled by Deep Denoising
4. Steepest Descent Density Control for Compact 3D Gaussian Splatting
5. Adaptive Denoising via GainTuning
6. Unsupervised Deep Video Denoising
7. Robust and Interpretable Blind Image Denoising via Bias-Free Convolutional Networks
Patent: Cascaded signal selection
Research Experience
Held research positions at Google, IST Austria, and the Flatiron Institute. Developed an AI-based denoising method for electron microscopy images in collaboration with materials scientists, revealing previously unobserved atomic-scale dynamics.
Education
Ph.D. from the Center for Data Science at New York University, advised by Prof. Carlos Fernandez-Granda and Prof. Eero P. Simoncelli; B.Tech in Electrical Engineering from the Indian Institute of Technology Madras, working with Prof. Kaushik Mitra.
Background
Staff Research Scientist at Meta AI, working at the intersection of Generative AI and 3D world perception, understanding, and creation. Previously, co-developed a photorealistic “AI twin” system for video calling, featured in Mark Zuckerberg’s Connect 2024 demo. Broader work spans image, video, and 3D content generation, multimodal systems, and computational photography. Has a deep interest in using AI to push the boundaries of scientific discovery.