Research Scientist, HRI Research to Enable Collaborative Humanoid Robots

DeepMind•New York, NY

12d•$141,000 - $202,000•Onsite

About The Position

We are looking for Research Scientists to join the Robotics team whose mission is to “Build embodied AI responsibly to benefit people in the physical world. This role is focused on foundational collaborative robotics research and unlocking capabilities to enable safe and effective interaction between people and robots. Research Scientists work on a diverse and stimulating range of projects including: inventing algorithms and prototype applications, working with real robots in the lab and outside the lab tackling real world use cases. Research Scientists work in large collaborative teams to make cutting edge breakthroughs to advance Robotics AI. We believe large foundation models will transform general purpose Robotics like with our Gemini Robotics models. This role will involve working with a diverse multi-functional team that builds the future of robotics.

Requirements

PhD in a technical field or equivalent practical experience.
Knowledge of the latest in large machine learning research.
Experience working with real-world robots
Expertise in using large datasets with deep neural networks to make real robots useful

Nice To Haves

Research track record in one or more of the following topics: HRI, Humanoid Whole Body Control, Vision Language Action models; Motion Planning, Force Control, AI Safety, Diffusion Policies, World Models, Imitation Learning and Reinforcement Learning, Sim2Real Transfer, Alignment Techniques

Responsibilities

Design, implement, train and evaluate large models and algorithms for humanoid robots.
Make breakthroughs and unlock new humanoid safety capabilities. Areas of focus include: Human Robot Communication: Verbal and Nonverbal, includes research on Bidirectional Native Audio Robotics Models, voice and persona development, and gesture generation.
Human Perception and Intent Understanding: Understanding where people are in the environment and what they are doing so that the robot can react appropriately.
Human Physical Safety Aware VLAs: Make VLAs safe with Safe Stops, Avoidance of unintentional collisions, compliant controllers, and behavior that modulates with respect to nearby people.
Write software to implement research ideas and iterate quickly.
Leverage your expertise to participate in a wide variety of research: learning from simulation, reinforcement learning, learning from demonstrations, vision-language-action models, transformers, video generation, robot control, humanoid robots and more.
Work effectively with a large collaborative team with fast-paced agendas to meet ambitious research goals.
Generate creative ideas, set up experiments and test hypotheses. Report and present research findings clearly and efficiently both internally and externally.