Principal Research Scientist, Robot Foundation Model
Core
Building foundation models for general-purpose robots (mobile manipulators, dual-arm, humanoids) to perceive, reason, move, and manipulate the physical world.
Role type
Principal Research Scientist (Robot Foundation Models)
Builds
Scalable data flywheels, new model architectures, learning algorithms, large models, and deployed robot policies.
Domain
Robotics + Foundation Models + Embodied AI
Deliverable
production ML models
Required skills
Machine learning (vision-language, video, robot policies, foundation models), scalable training (multi-node, large datasets/models), research track record (top-tier publications), coding (modern ML frameworks), translating research to working systems, data strategy, model design
Preferred skills
PhD in CS/ML/Robotics/CV, 5+ years industry experience, real robot/embodied AI/simulation experience, large-scale video data, sequential decision-making, mentoring
Technologies
Vision-language-action models, world and action models, omni-modal models, video generation, reinforcement learning, imitation learning, behavioral cloning, distributed training infrastructure
Responsibilities
Lead research on robot foundation model architectures and learning approaches; Design and evaluate VLAs, WAMs, and omni-modal models; Develop RL, behavioral cloning, and policy learning methods; Curate large-scale video datasets; Build distributed training pipelines; Influence technical decisions on policy and data strategy; Collaborate with engineering and robotics teams; Contribute to external publications
Seniority
Principal, 0→1 ownership & strategy
