Job Description

Minimum qualifications

• Bachelor’s degree or equivalent practical experience.
• 5 years of experience with software development in one or more programming languages.
• 3 years of experience testing, maintaining, or launching software products, and 1 year of experience with software design and architecture.
• Experience with ML frameworks such as JAX, TensorFlow, or PyTorch.

Preferred qualifications

• Master's degree or PhD in Computer Science or related technical field.
• 1 year of experience in a technical leadership role.
• Experience building agents for 3D virtual environments, simulators, or video games.
• Strong track record in machine learning, data science, or AI in games, which can include publications in conferences (NeurIPS, ICLR, ICML, CVPR, etc.).
• Domain knowledge in game development, concepting, or community.
• Solid understanding of LLM internals, (e.g., typical training pipelines, computational characteristics of training/inference, etc.), and knowledge of Deep Reinforcement Learning (RL), LLM Reasoning, imitation learning, memory-based architectures, Vision-Language-Model (VLM), or Vision-Language-Action (VLA) models.

About the job

DeepMind has long-established connections with gaming, whether through early AI-breakthroughs mastering Atari, Go, StarCraft, and more recent research with SIMA agents. We are seeking a highly motivated and innovative Research Engineer to join our team in Tokyo, focused on building multimodal embodied gaming agents.

In this role, you will work with researchers and engineers to develop embodied gaming agents capable of perceiving, reasoning, planning, and executing precise real-time actions in complex, open-ended environments. You will utilize the latest advancements in multimodal large language models (LLMs), vision-language-action (VLA) models, in-context learning (ICL), supervised fine-tuning (SFT), and reinforcement learning (RL) to solve fundamental challenges in embodied intelligence. You will leverage these architectures to bridge the gap between high-level long-horizon planning and low-level high-frequency motor control, creating agents that can adaptively master tasks in rich virtual testbeds (including high-fidelity 3D simulations and sandbox games).

This role offers you a unique opportunity to stand at the forefront of the quest for AI. You will join an exceptional team solving the "hard problems" of embodiment—autonomously solving long-horizon tasks, learning from vast multimodal memories, and generalizing to completely unseen worlds. If you are passionate about pushing the frontiers of what AI agents can achieve and are eager to define the next era of adaptive intelligence, we encourage you to apply.

Artificial intelligence will be one of humanity’s most transformative inventions. At Google DeepMind, we are a pioneering AI lab with exceptional interdisciplinary teams focused on advancing AI development to solve complex global challenges and accelerate high-quality product innovation for billions of users. We use our technologies for widespread public benefit and scientific discovery, ensuring safety and ethics are always our highest priority.

We are pushing the boundaries across multiple domains. Our global teams offer diverse learning opportunities and varied career pathways for those driven to achieve exceptional results through collective effort.

We are pushing the boundaries across multiple domains. Our global teams offer diverse learning opportunities and varied career pathways for those driven to achieve exceptional results through collective effort.

Responsibilities

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