Shivam Singhal

University of Washington, Tata Consultancy Services (India)

Papers

Total Citations

H-Index

About

Shivam Singhal’s research bridges two transformative frontiers: dexterous robotic manipulation and infrastructure-free 3D tracking. His most influential work, “Learning Deep Visuomotor Policies for Dexterous Hand Manipulation” (43 citations), pioneered the use of deep reinforcement learning to enable multi-fingered hands to perform complex, real-world skills—grasping, in-hand manipulation, and tool use—using only on-board visual sensing. This approach moves beyond traditional, sensor-heavy setups, offering a scalable path toward truly versatile robotic hands. Complementing this, his paper “AiRite” (6 citations) tackles the challenge of precise 3D tracking for smart wearables without relying on external infrastructure like cameras or acoustic arrays. By developing a sensor-agnostic approximation model, Singhal’s work supports applications in augmented reality, gesture recognition, and indoor localization. Together, these contributions demonstrate a rare ability to advance both the physical intelligence of robots and the perceptual capabilities of everyday devices. Singhal’s research is characterized by its practical ambition—seeking solutions that work in the messy, infrastructure-poor real world—making him a notable emerging voice in embodied AI and ubiquitous computing.