Research Scientist, Robotics, Cloud AI

• Develop novel techniques for robotic grasping, dexterous manipulation, planning, navigation, and 3D spatial understanding in dynamic physical environments.
• Design scalable teleoperation systems, portable data capture interfaces, and agent-assisted workflows to drive robot learning.
• Create advanced sim-to-real frameworks to enable scalable, generalized policy learning.
• Architect, test, and deploy robust hardware-software robotic systems to automate complex, high-precision physical workflows under real-world constraints.
• Partner with researchers, hardware specialists, and engineering teams to push the frontiers of physical AI across Google.

Minimum qualifications:

• PhD in Computer Science or robotics related fields, or equivalent practical experience.
• 2 years of experience in research across robotics domains, such as navigation, locomotion, manipulation, data collection for robot learning, sim-to-real transfer, vision-language-action model, or world action model development.
• One of more scientific publication submission(s) for conferences, journals, or public repositories (NeurIPS, ICML, ICLR, CoRL, RSS, ICRA, IROS, CVPR, ECCV, ICCV, etc.
• Experience coding in Python or C++.

Preferred qualifications:

• 2 years of coding experience in Python or C++, as well as software frameworks such as PyTorch, JAX, ROS2, OpenCV.
• 1 year of experience owning and initiating research agendas.
• Experience building and scaling real-world robotic systems end-to-end, spanning custom data collection pipelines, model training, and physical deployment.
• Demonstrated ability to bridge theoretical algorithms with physical robotic hardware or high-fidelity simulation environments.

• PhD in Computer Science or robotics related fields, or equivalent practical experience.
• 2 years of experience in research across robotics domains, such as navigation, locomotion, manipulation, data collection for robot learning, sim-to-real transfer, vision-language-action model, or world action model development.
• One of more scientific publication submission(s) for conferences, journals, or public repositories (NeurIPS, ICML, ICLR, CoRL, RSS, ICRA, IROS, CVPR, ECCV, ICCV, etc.
• Experience coding in Python or C++.