We're looking for a Research Scientist or Research Engineer to advance dexterous manipulation - enabling our robots to perform contact-rich, fine-motor tasks that require precision, physical reasoning, and adaptability to novel objects and environments. What You'll Do • Research and develop learning-based approaches for dexterous and contact-rich manipulation tasks • Design training strategies and data collection protocols for fine-motor and multi-finger manipulation • Work on perception for manipulation: contact detection, tactile sensing, object pose estimation, and spatial reasoning • Build and evaluate policies that generalize to novel objects and unstructured environments • Develop simulation environments and benchmarks for dexterous manipulation research • Collaborate with robot hardware, perception, and learning teams to close the sim-to-real gap • Publish and present work at top-tier robotics and ML venues (especially valued for RS track) What We're Looking For • Strong background in robot learning, manipulation, or physical AI • Hands-on experience developing and evaluating manipulation policies on real hardware • Understanding of contact mechanics, grasp planning, or tactile sensing • Solid ML skills with experience in imitation learning, RL, or diffusion-based policies • Ability to work across the stack from simulation to real robot deployment Nice to Have (But Not Required) • PhD in Robotics, ML, or a related field • Publication record at ICRA, CoRL, RSS, NeurIPS, or related venues • Prior work on dexterous hands, multi-finger manipulation, or contact-rich tasks • Experience with tactile sensors or force/torque feedback in robot learning • Familiarity with simulation tools for manipulation (MuJoCo, Isaac Sim, Genesis) • Experience with skill libraries, language-conditioned manipulation, or task parameterization Why This Role • Push the frontier on one of the hardest open problems in robotics • Work with hardware and data resources that few research labs have access to • Direct path from research results to deployment on our humanoid platform • Tight collaboration across robot learning, hardware, and systems teams