Chris Paxton

Robotics researcher · Robot learning and mobile manipulation

Pittsburgh, PA  ·  chris.paxton.cs@gmail.com  ·  cpaxton.github.io  ·  GitHub  ·  Scholar

I build robots that can find and manipulate objects in unfamiliar spaces. My work spans humanoid policy training, open-source systems for home robots, and benchmarks that let research teams test the same task in simulation and on hardware.

Experience

2025–presentAgility Robotics · Principal AI Scientist

Lead a team of researchers and engineers working on whole-body reinforcement learning and imitation learning for humanoids. Built the AWS training infrastructure, from data processing to distributed policy training, and set the team’s research priorities.

2024–2025Hello Robot · Senior Lead, Embodied AI

Led development of Stretch AI, an open-source system that takes a spoken instruction, finds the requested object, and guides a mobile robot to pick it up. Integrated open-vocabulary perception, spatial memory, and policies learned from demonstrations.

2022–2024Meta FAIR · AI Research Scientist

Led a collaboration of 15+ researchers at Meta, Carnegie Mellon, and Georgia Tech on HomeRobot. Built a shared benchmark and robot software stack for open-vocabulary mobile manipulation and launched the NeurIPS 2023 challenge.

2018–2022NVIDIA · Seattle Robotics Lab

Advanced from postdoctoral researcher to Senior Robotics Research Scientist. Developed learned perception and reactive planning methods for grasping in clutter and sequential manipulation. Mentored five or more graduate student projects.

2016–2017Zoox · Research Engineer Co-op

Worked on motion planning and deep learning for autonomous vehicles.

Selected research and public work

OK-Robot · Co-authored a zero-shot mobile manipulation system. Across ten real homes it completed 58.5% of full pick-and-drop tasks.

6-DOF Grasping for Target-driven Object Manipulation in Clutter · Co-authored work on planning grasps from partial point clouds. ICRA 2020 finalist for best manipulation and student paper.

CoSTAR · Led the Johns Hopkins team that won the 2016 KUKA Innovation Award for a system that lets people teach industrial robots tasks.

RoboPapers · Interview researchers about how their systems work and what their experiments show. Write about robotics and AI at It Can Think.

Education

PhD, Computer Science · Johns Hopkins University, 2018. Advisor: Gregory D. Hager.
Creating Task Plans for Collaborative Robots

BS, Computer Science · University of Maryland, College Park, 2012. Minor in Neuroscience.