Noam Shazeer

Noam Shazeer is an American computer scientist and entrepreneur known for inventing the Transformer architecture and sparsely-gated Mixture of Experts, the foundations of modern large language models. Born in Philadelphia and raised in Massachusetts, he won a gold medal with a perfect score at the 1994 International Mathematical Olympiad as a member of the US team before studying math and computer science at Duke University. He joined Google in 2000, where he worked for two decades and co-authored 'Attention Is All You Need' (2017). In 2021 he left Google to co-found Character.AI and served as its CEO; Google paid about $2.7 billion in 2024 to license its technology and bring him back as VP Engineering and co-tech-lead of Gemini. In June 2026 he left Google to join OpenAI.

Notes5

• Left Google for OpenAI in June 2026, less than two years after the $2.7B return deal, suggesting the Gemini co-lead role was not enough to retain him.Jun 18, 2026
• Left Google in 2021 to co-found Character.AI and serve as CEO; Google then paid ~$2.7B in Aug 2024 to license the technology and rehire him, a reverse-acquihire that returned him to Google as VP Engineering and Gemini co-tech-lead.Aug 2024
• Personally designed the multi-head attention and residual architecture of the Transformer and coded its first better-than-SOTA implementation, making him the practical author of the architecture, not just a co-signer.Jun 2017
• Won a gold medal with a perfect score at IMO 1994 representing the USA, the mathematical-prodigy origin that predates his AI career.Jul 1994
• Joined Google in 2000 and wrote the PHIL algorithm that became the core of Google AdSense, so his commercial impact at Google predates deep learning by a decade.
• Elected to the National Academy of Engineering in 2026, per the NAE class announcement referenced by his Wikipedia article.Feb 10, 2026
• Named to the TIME100 AI list in 2023 as CEO and co-founder of Character.AI, recognition that anchored his public profile before the Google deal.Sep 7, 2023