Rajiv Mathews is a Distinguished Software Engineer at Google DeepMind where he is the DRI for Gemini Live engineering. His specific focus is on aspects of large-language modeling, in particular multimodality, personalization, agentic behaviors, and real-time interactivity. Previously he led groups working on privacy-preserving machine learning techniques, focused on federated learning, differential privacy and on-device personalization, with applications in natural language, speech and mobile keyboards.

Scholar

Google Research

LinkedIn