Applied Research - Robotics
Core
Build environments, evaluations, agent systems, and post-training methods for embodied AI and robotics, training agents with reinforcement learning and multimodal perception using simulation and real-world robot feedback.
Role type
Applied Research Scientist (Robotics & Embodied AI)
Builds
Simulation and real-world environments, long-horizon agents, training loops combining offline/online RL and demonstrations, and reward/evaluation systems.
Domain
Robotics, Embodied AI, Reinforcement Learning, Computer Vision
Deliverable
production ML models
Required skills
Robotics, Embodied AI, Reinforcement Learning, Imitation Learning, Control Systems, World Modeling, Simulation, Computer Vision, Agent Evaluation, Python
Preferred skills
Experience with production systems interacting with the physical world, strong experimental judgment, ability to build extensible artifacts
Technologies
Python, Reinforcement Learning frameworks, Simulation environments
Responsibilities
Build aligned simulation and real-world evaluations for long-horizon embodied tasks; Combine demonstrations, interventions, offline data, and online feedback into training loops; Train embodied agents to recover from partial failures and changing scenes; Design rewards and evaluations for task success, safety, and continual improvement.
Seniority
Mid-Senior, hands-on IC
