The Staff Research Scientist role at DeepMind focuses on advancing machine learning by inventing new algorithms and training methods. The position requires a deep understanding of ML architectures and the ability to evaluate AI-generated ideas. As a key technical verifier, the scientist will influence the direction of research and ensure the implementation of validated improvements in production models. This role is integral to the Gemini Research team, which aims to push the boundaries of AI technology for global benefit.

• PhD in Computer Science, a related field, or equivalent practical experience.
• 4 years of experience with research agendas across multiple teams or projects.
• Experience working with deep learning frameworks such as JAX/Flax or PyTorch for distributed systems and accelerator clusters.
• One or more scientific publication submission(s) for conferences, journals, or public repositories.

• Invent, implement, and scale new machine learning algorithms, including architectures and optimizers.
• Design and build automated research systems for Gemini to propose and evaluate changes to its training stack.
• Serve as the final verifier for AI-proposed algorithm changes and reproduce results.
• Build and maintain high-performance training and evaluation code to allow testing of ideas at scale.
• Collaborate with pre-training, post-training, and infrastructure teams to implement validated improvements.

• Employees at Google are often offered benefits like comprehensive health insurance, 401(k) matching, and flexible work arrangements, among other benefits.