Jeff Dean

If modern AI has a technical backbone, a big part of it was built by Jeff Dean. He joined Google in 1999, when the company had just a few hundred employees, and set out to solve a problem nobody had cracked before: how to make thousands of cheap computers work together as one. That invisible work nobody sees is the reason Google can index the entire internet.

In 2004 he co-invented MapReduce, the framework that revolutionized distributed computing, and in 2005 he co-designed BigTable, the database that still powers Gmail, YouTube, and Search today. But his boldest move came in 2011: he co-founded Google Brain, the deep neural network research team, at a time when neural networks were still seen as an academic curiosity, not production technology. In 2015, Google Brain launched TensorFlow, which became the world's most widely used deep learning library for years. And to make all of that run fast, Dean led the team that designed TPUs, the custom chips Google built from scratch to accelerate AI.

That work earned him a list of recognitions that speaks for itself: the ACM Prize in Computing in 2012 alongside his longtime collaborator Sanjay Ghemawat, the IEEE John von Neumann Medal in 2021, membership in the National Academy of Engineering since 2009 and in the American Academy of Arts and Sciences since 2016. In 2023 he was named Chief Scientist of Google DeepMind, the unit that merged Google Brain with the legendary DeepMind lab. And in August 2026, after 27 years at Google, he surprised the world by announcing he was leaving to co-found Discovery Loop, a startup focused on accelerating scientific research with AI.

What strikes me most about Jeff Dean isn't the award list or the systems he built: it's the consistency. From the moment he arrived at a small company, he bet on building things others considered impossible or unnecessary. First distributed computing, then neural networks at scale. Every time you use a Google app, a TensorFlow model, or an AI chip, you're touching a little piece of that bet. When someone with that track record decides to start over at 58, it's worth paying attention.