Research Engineer, Embodied Generalist Agent

As a Research Engineer at Google DeepMind, you will contribute to the development of Gemini-powered embodied agents capable of autonomous progression and complex problem-solving.

We believe that rich virtual environments provide the ideal pressures to develop robust skills in reasoning, memory, and motor control. You will use these domains to research how agents can learn from demonstrations and experiences, and adapt their strategies in real-time.

Key responsibilities

• Agent Architecture & Control: Develop and optimize state-of-the-art agent architectures that seamlessly integrate multimodal perception, reasoning, and precise real-time execution.
• Scalable Learning: Build and scale training recipes utilizing supervised fine-tuning, reinforcement learning, imitation learning, and/or in-context learning..
• Memory & Planning: Design advanced systems that enable agents to reason over long horizons and effectively utilize memory to solve complex, extended tasks.
• Adaptation: Research and implement capabilities that allow agents to adapt to new environments and learn from experience at test time.
• Evaluation: Establish rigorous benchmarks within virtual environments to measure progress in general agent capabilities and embodied intelligence in unseen environments.

• Bachelors/Masters/Ph.D. in Computer Science, Artificial Intelligence, or a related field.
• Experience with relevant ML frameworks such as JAX, TensorFlow, or PyTorch.
• Strong programming skills in Python and experience with large-scale data pipelines.
• Solid understanding of LLM internals, e.g., typical training pipelines, computational characteristics of training/inference, mechanisms for multimodal extension.
• Knowledge of Deep Reinforcement Learning (RL), LLM Reasoning, Imitation Learning, Memory-Based Architectures, Vision-Language-Model (VLM), and/or Vision-Language-Action (VLA) models.
• Proven track record of designing, implementing, and maintaining robust technical assets (such as libraries, frameworks, or models) used by a large number of technical stakeholders; experience with OSS contributions is a plus.
• Excellent communication and collaboration skills.

• A minimum of 5 years of relevant professional experience.
• Experience building agents for 3D virtual environments, simulators, or video games.
• Strong track record in competitions in machine learning, data science, or AI in games.
• Strong track record in AI competitions or publications in top-tier conferences (NeurIPS, ICLR, ICML, CVPR, etc.).