I’m a Research Scientist at Dolby Laboratories, India. At Dolby, my research focuses on enabling native HDR Video/Image generation in Foundational Models, Sparse-View 3D reconstruction to Gaussian Splats and Multimodal Splat Editing. I completed my PhD in Computer Science from University College London advised by Prof. Niloy J Mitra, as part of the Smart Geometry Group. My thesis: Towards 3D Generative Models With Sparse Guidance presents novel learning algorithms that enable neural 3D pose-shape generation from super-sparse data examples (less than 100) for both static and animated 3D content. Before my PhD, I worked as a Research Engineer at Adobe Research’s Creative Intelligence Lab, where my research was advised by Siddhartha Chaudhuri and focused on novel neural shape representations and learning frameworks to enhance discriminative tasks such as segmentation, correspondence and shape retrieval. I completed my Master’s in Computer Science at IIT Bombay, advised by Prof. Siddhartha Chaudhuri, and covered by Qualcomm Innovation Fellowship.

Research

High-fidelity content generation across modalities relies on powerful neural representations that disentangle underlying factors and incorporate strong physical priors. My research focuses on developing these disentangled representations and physical priors for 3D scenes, images, and videos. By doing so, I aim to enable data-efficient, intuitive workflows for visual media creation.

Publications