Staff Research Scientist/Engineer, Consumer Agents Post Training, DeepMind

Job description

info_outline About the job This role focuses on deep learning, LLMs, and reinforcement learning. Because of this focus, candidates with primarily ML infrastructure (ML infra) backgrounds will likely find this is not the right profile for their skill set. We are looking for a Research Engineer/Research Scientist to work on Agent Post-training and Personalization to build Consumer Agents. You will research, develop and scale foundational post-training methods that convert foundation models into autonomous, adaptive agents for consumers.

Your focus centers on reinforcement learning, targeted data creation, and persistent context architectures so models retain user-specific state reliably over time and adapt to individual users. Additionally, you will build training datasets and evaluation frameworks to test long-horizon reasoning, uncover failure modes, and measure agent reliability across complex workflows.