Group Product Manager, Architectural Advances, Frontier AI, DeepMind

Google•London, England

17h•$240,000 - $333,000

About The Position

As a Product Manager at DeepMind, you will be responsible for shipping AI research breakthroughs into Google’s AI models and products. You will work with AI researchers advancing model architectures, AGI, and applications such as agents and robotics. The role involves understanding how new architectures affect Gemini and other models, helping teams navigate tradeoffs, and collaborating with model development teams. You will work with product areas such as Cloud and Search to understand customer needs and how model advances can address them. You may ship architectural advances into existing models, introduce new model families, or work with product teams to ship custom models. You will also help research teams develop communication and ecosystem strategies to support awareness and adoption of model advancements. DeepMind is a pioneering AI lab with interdisciplinary teams working to advance AI, solve complex global challenges, and support product innovation. The organization emphasizes widespread public benefit, scientific discovery, safety, and ethics.

Requirements

Bachelor's degree in Computer Science or equivalent practical experience.
10 years of experience building AI products, including AI architecture and models.

Nice To Haves

Understanding of the current model landscape and the strengths and weaknesses of various techniques.
Ability to stay up to speed on relevant research in core model advancements.
Ability to anticipate model gaps that will need to be addressed to reach AGI.
Ability to develop clear goals and strategies for turning model advancements into product features.
Strong ability to communicate complex technical advancements simply.
Strong ability to collaborate in high-stakes environments with multiple executive stakeholders and technical leads.

Responsibilities

Understand the implications of new model architecture breakthroughs for Gemini and other models.
Build collaborations with product teams and model development teams to ship model architecture advances.
Work with Google product areas to understand customer needs that specific model architecture advancements could address more effectively.
Build an ecosystem of tooling and infrastructure that enables new model advances to become viable alternatives to existing ones.
Deploy entirely new models.
Help research teams build communication and ecosystem strategies for model advancements to encourage adoption and awareness.