Summary
Tom Zahavy is a research scientist at DeepMind with nine years of experience specializing in deep reinforcement learning and creative agent development, contributing to projects like AlphaProof, AlphaZero_db, PuzzleGen, and Convex RL. He holds a Ph.D. track from Technion where his background in electrical engineering and physics underpins a strong blend of theory, principled experimentation, and applied ML. His trajectory includes research internships at Google, Facebook, Microsoft, and Walmart eCommerce, giving him broad exposure to industry-scale problems from 3D sensing to transfer learning and domain adaptation. At DeepMind he focuses on pushing algorithmic frontiers for autonomous problem-solving rather than just engineering pipelines, often bridging ideas from generative methods and RL. Based in London, he pairs academic rigor with product-minded research, shipping tools and datasets that support large-scale agent evaluation and reproducible science. Colleagues describe him as someone who turns elegant theory into practical systems that reveal surprising capabilities in agents.
9 years of coding experience
3 years of employment as a software developer
Bachelor of Science (B.Sc.), EE + Physics (double degree), Bachelor of Science (B.Sc.), EE + Physics (double degree) at Technion - Israel Institute of Technology
English, Spanish, Hebrew