Member of Technical Staff, Robotics & Embodied AI Research

What You'll Do7

• Drive frontier research in VLA models, generalist policies, world models, and new embodied AI methods.
• Build end-to-end experimentation systems that cover teleoperation, data collection, simulation, policy deployment, and real-world evaluation.
• Create high-fidelity simulations from real-world data to assess robot policies and confirm performance in physical settings.
• Use flow, a human-in-the-loop annotation platform, to organize and scale robotics datasets from egocentric video, teleoperation demonstrations, and multimodal sensor data.
• Probe model limitations, failure modes, and sim-to-real gaps, then design better datasets, benchmarks, and evaluation methods.
• Prototype and validate new approaches to robot learning, policy optimization, and embodied intelligence at a fast pace.
• Find high-impact research opportunities on your own and lead projects from early exploration through experimentation and scalable implementation.

Requirements8

• Solid background in robotics, embodied AI, robot learning, computer vision, or applied machine learning research.
• Hands-on experience developing, training, or evaluating robot learning models, including VLA models, imitation learning, or generalist policies.
• Proficiency in Python and experience with robotics or ML frameworks such as PyTorch, ROS, or simulation environments.
• Experience designing robotics experiments that involve real-world data collection, simulation, policy evaluation, or hardware deployment.
• Understanding of robot learning methodologies, multimodal perception, and sim-to-real transfer challenges.
• Ability to analyze model behavior, identify failure modes, and turn research findings into measurable performance gains.
• Strong research ownership, experimental judgment, and the ability to execute complex technical projects independently.
• Strong technical communication and collaboration skills across research and engineering teams.

Preferred Qualifications6

• Experience with world models, reinforcement learning, imitation learning, or diffusion-based robot policies.
• Hands-on experience with teleoperation, robotics hardware, sensor integration, or real-world policy deployment.
• Familiarity with robotics simulators such as MuJoCo, Isaac Sim, or comparable environments.
• Experience building multimodal robotics datasets from egocentric video, robot trajectories, or sensor data.
• Prior experience at a leading robotics or frontier AI research lab.
• Contributions to robotics research publications, open-source frameworks, or benchmark development.

Who Should Apply

This role suits a researcher with a strong robotics or embodied AI background who has trained or evaluated VLA models, imitation learning, or generalist policies. You should be comfortable with Python, PyTorch or ROS, and simulation tools, and you have designed experiments that connect simulation to real-world robot deployment. You work independently, spot failure modes, and turn findings into measurable improvements.

Salary Insight