Research Scientist, Robotics Pre-Training and Data Quality, DeepMind

Overview

As a Research Scientist, you drive data quality, acquisition, and labeling strategies for robotics foundation models and oversee data-driven mixture optimization. You will coordinate training runs, refine training recipes, and build tools to accelerate research iteration. You contribute to RL, vision-language-action modeling, world-action models, and simulation within a fast-moving, collaborative team. You’ll push ideas from lab to robust real-world robotic systems while prioritizing safety and broad impact.

Responsibilities

• Drive data quality, acquisition, and labeling strategies for robotics foundation models
• Oversee data mixture optimization for robotics foundation models
• Coordinate training runs and iterate on training recipes
• Build and maintain software tools and processes to enable rapid research iteration
• Verify data quality through robot policy training and imitation learning workflows
• Contribute to broader research areas including reinforcement learning, vision-language-action modeling, world-action models, and simulation
• Set up large-scale tests, deploy promising ideas, and manage deadlines and deliverables
• Publish findings and share research insights with the wider community
• Advance real-world robotic systems by bridging lab research with practical deployment

Key requirements

• PhD in a technical field or equivalent practical experience
• Experience with algorithmic architectures, data sources, and training/inference techniques for generative multimodal models in robotics (e.g., VLAs, WAMs)
• Experience optimizing models and systems (performance tuning, experimentation, debugging)

• collaborative mindset
• problem-solving orientation
• ability to manage and communicate complex projects
• generative multimodal models for robotics (VLAs, WAMs)
• model and system optimization (performance tuning, experimentation)
• training recipe design and benchmarking

…