I am currently a Principal Research Scientist at NVIDIA Robotics and co-lead the Dexterity robotics research team. Prior to that, I was a Research Scientist at OpenAI and before that I obtained my PhD from Imperial College London.
Bio
I began my career in hardware and robotics, completed a PhD in computer vision, and then returned to robotics with a stronger foundation in machine learning and vision techniques. My work now heavily relies on simulation and after exploring the role that rendering engines can play in generating data for understanding and building SLAM systems, I have been combining physics simulation with rendering to develop robust multi-fingered dexterous manipulation systems. This approach has powered my recent Dex series of work over the past several years.
Selected Publications
-
1.
Isaac Lab: A GPU-accelerated simulation framework for multi-modal robot learningproject website / arxiv
-
2.
End-to-End RL Improves Dexterous Grasping Policiesproject website / arxiv
-
3.
CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Modelsproject website / arxiv
-
4.
DextrAH-RGB: Visuomotor Policies to Grasp Anything with Dexterous Handsproject website / arxiv
-
5.
DextrAH-G: Pixels-to-Action Dexterous Arm-Hand Grasping with Geometric Fabricsproject website / arxiv
-
6.
DeXtreme: Transfer of Agile In-Hand Manipulation from Simulation to Realityproject website / arxiv
-
7.
CuRobo: Parallelized Collision-Free Robot Motion Generationproject website / arxiv
-
8.
DexPBT: Scaling up Dexterous Manipulation for Hand-Arm Systems with Population-Based Trainingproject website / arxiv
-
9.
Transferring Dexterous Manipulation from GPU Simulation to a Remote Real-World TriFingerproject website / arxiv
-
10.
Isaac Gym: High Performance GPU-Based Physics Simulationproject website / arxiv
-
11.
DexPilot: Vision-Based Teleoperation of Dexterous Robotic Hand-Arm Systemproject website / arxiv
-
12.
Robust Learning of Tactile Force Estimation through Robot Interactionproject website / arxiv
-
13.
Closing the Sim-to-Real Loop: Adapting Simulation Randomization with Real World Experienceproject website / arxiv
-
14.
GPU-Accelerated Robotic Simulation for Distributed Reinforcement Learningproject website / arxiv
-
15.
SceneNet RGB-D Segmentation: Can 5M Synthetic Images Beat Generic ImageNet Pre-training on Indoor Segmentation?project website / arxiv
-
16.
Understanding Real World Indoor Scenes With Synthetic Dataproject website / arxiv
-
17.
A Benchmark for RGB-D Visual Odometry, 3D Reconstruction and SLAMproject website
-
18.
Simultaneous Mosaicing and Tracking with an Event Cameraproject website
-
19.
Real-Time Camera Tracking: When is High Frame-Rate Best?project website