Purva Tendulkar

Papers

Total Citations

H-Index

About

Purva Tendulkar is a researcher advancing the intersection of computer vision, graphics, and robotics, with a focus on generating realistic, interactive 3D human avatars. Her most notable contribution is the development of **FLEX**, a pioneering method for full-body grasping that synthesizes 3D human avatars—including hands and entire bodies—interacting naturally with everyday objects. This work addresses a critical challenge in AR/VR, video games, and robotics, achieving 30 citations since 2023 by enabling realistic scene interaction without requiring expensive full-body grasp data. Tendulkar also tackles fundamental machine learning problems, as seen in her work on **Landscape Learning for Neural Network Inversion**, which improves the stability and efficiency of gradient-based inversion methods for solving inverse problems in vision and graphics. Her research demonstrates a rare ability to bridge practical application needs with core algorithmic innovation, making her a rising voice in embodied AI and human-scene interaction. With her focus on making virtual humans move and grasp more naturally, Tendulkar’s work is poised to influence how we design immersive digital experiences and train robots to interact with the physical world.