Research Scientist, Multisensor Robotics, DeepMind

Overview

As a Research Scientist at Google DeepMind, you will architect and train models to expand robotic capabilities and prototype concepts, bridging theory and real-world deployment. You will tackle sim-to-real transfer, imitation learning, and multimodal architectures, contributing to embodied AI innovations. You’ll set up large-scale tests, deploy promising ideas, and communicate results to diverse audiences. This role offers the chance to impact real-world robotics with cutting-edge research and global collaboration.

Responsibilities

• Architect and train models to expand robotic capabilities and prototype concepts
• Advance sim-to-real transfer, imitation learning, Vision-Language-Action (VLA) models, and multi-sensory learning
• Propose novel ideas and address manipulation hardware requirements
• Communicate results to audiences and stakeholders
• Set up large-scale tests and rapidly deploy promising ideas
• Contribute to the wider research community via publishing and collaborations
• Bridge theoretical research and real-world deployment at scale in embodied AI
• Develop intelligent systems for meaningful physical world interaction

Key requirements

• PhD in a relevant quantitative field or equivalent practical experience
• 5 years of professional research experience in an industrial R&D lab
• 3 years designing, training, and scaling RL, IL, or multimodal generative architectures (including VLA or video foundation models)
• 2 years optimizing foundation models for production (distillation, pruning, post-training quantization) for on-device latency and memory constraints

• strong communication and presentation skills
• ability to manage deadlines and deliverables
• cross-functional collaboration
• reinforcement learning (RL)
• imitation learning (IL)
• multimodal generative architectures (including Vision-Language-Action, video foundation models)

…