Research Scientist (Robotics) - Humanoid Robot Data Collection

About the role

The AI Robot Association (AIRoA) is launching a project to collect one million hours of humanoid robot operation data with hundreds of robots, using it to train powerful Vision-Language-Action (VLA) models. The project is a major national initiative supported by 20.5 billion yen from NEDO, and AIRoA aims to make its datasets, models and benchmarks open to everyone. This Research Scientist (Robotics) role is based in Tokyo.

Key responsibilities

• Design and build scalable data collection systems operating across hundreds of humanoid and mobile robots.
• Develop semi-autonomous teleoperation systems with low latency, high controllability and minimal operator workload.
• Create automated pipelines to deploy trained models and evaluate them in simulation and on real robots.
• Work closely with the VLA team to integrate and deploy models on physical robots.

Required profile

• Ph.D. in Robotics, Mechanical Engineering, Computer Science or a related field, or M.S. with 3+ years of equivalent industry or research experience.
• Extensive hands-on experience with physical robotic platforms such as mobile robots, manipulators (UR5, Franka Emika, HSR) and humanoids (Spot).
• Ability to control real robots with ROS / ROS2, including full-stack integration and deployment.
• Solid understanding of teleoperation, AR/VR interfaces and leader-follower architectures.
• Strong C++ and Python, with production-level code and system-level integration.
• Proven ability to work independently and collaboratively in multidisciplinary teams; purely theoretical candidates (e.g. reinforcement learning, explainable AI) will not be considered.

Preferred qualifications

• Papers in RSS, CoRL, ICRA, IROS, TRO, IJRR; CVPR, ICCV, NeurIPS publications may be scored lower.
• Advanced control on hardware (impedance, force/torque, admittance, MPC) and tactile sensing or contact-rich manipulation.
• Large-scale robot data collection or fleet management experience.
• Simulators (Isaac Sim, MuJoCo, PyBullet) and Sim2Real or VLA integration experience.