RAI Institute
Published:
Full-time · 3+ yrs Cambridge, Massachusetts, United States
Principal Research Lead
Sep 2023 – Present · On-site
Inventions & Technical Innovation
- Pioneered handheld force-based data collection system for robot training, capturing force, vision, and proprioception for in-the-wild demonstrations
- Built video prediction system for robot policies using internet-scale video as prior (early 2023, before the wave); requires only limited real-world demonstrations
- Designed multimodal architectures combining novel sensors with internet-scale priors for improved behavioral execution (publication pending)
- Developed gradient-free RL enabling online learning with non-differentiable semantic functions (U.S. Patent pending, May 2025)
- Created task definition framework that makes human demonstrations more learnable for robots
- Built ROS 2 interfaces and controllers for custom grippers leveraging novel force/torque sensors
Scaling Innovations
- Founded two research teams to scale innovations: Foundation Models (10+ researchers) and Capture (30 people)
- Capture is the largest data collection effort at the institute with a roadmap for 100,000+ demonstrations
- Built 320 H100 GPU infrastructure for large-scale training of video prediction and multimodal models
- Established research partnerships with Google, Columbia, ETH Zurich, Agile Robots
Research & Publications
- Published at CoRL 2024, IJCAI 2024, ICML 2023
- Co-authoring “Elephants Don’t Write Sonnets” (MIT Press 2026) with Stefanie Tellex
- Co-organized New England Manipulation Symposium (NEMS) 2025 at MIT
Research Scientist
Dec 2022 – Aug 2023 · 9 mos
- Joined to build foundation models research capability from scratch
- Invented early video prediction approach for robot policies leveraging internet-scale video data
- Pitched and launched Foundation Models team
- Created benchmark protocol for multiple teams based on state-of-the-art robotics manipulation datasets