Staff Software Engineer, GeminiApp Personalization, DeepMind

About the job

At Google DeepMind our mission is to build the world's first general-purpose learning agent. Central to this mission is the complex task of measuring the intelligence of our prototypes. As a Software Engineer, you will be working with the cutting edge AI agents developed by our exceptional team of Machine Learning and Neuroscience research scientists. Your responsibilities will include everything from creating systems for agent testing using 2D and 3D games to developing test problems within physics simulators. You will create graphical visualization of results, build competitive agent leaderboards and test new algorithms on robots. To succeed in this role you will need to have a strong foundation in software engineering and enjoy working on a wide range of challenging problems within a mission-driven team.

We are the Gemini App Personalization team at Google DeepMind, dedicated to building Google's next-generation AI assistant. Our mission is to empower billions of people by offering deeply personalized products that act as a seamless extension of their own intellect. In this role, you will help build a personal AI assistant that continuously absorbs, organizes, and effortlessly recalls the unique interests, passions, and curiosities of individuals, evolving alongside them to amplify their everyday thinking.

Responsibilities

• Design, prototype, and build scalable features on the full Gemini App stack that securely capture, organize, and intuitively surface long-term personal context.
• Perform data analysis of user feedback, logs, and evaluation tasks to identify opportunities for improving how effectively the assistant retains and synthesizes historical user interactions.
• Develop evaluation techniques (both automated and human-in-the-loop) to assess and hill-climb on the quality of contextual recall and the assistant's ability to act as a seamless cognitive extension.
• Act as the primary owner of model output quality, ensuring responses accurately reflect, synthesize, and build upon the user's unique history and ongoing preferences.
• Contribute to the development of a data flywheel that safely accumulates and leverages continuous user context, driving ongoing improvement and innovation.