Barret Zoph lands at Google after a year of bouncing between OpenAI and Thinking Machines
The AI researcher has held four senior roles in roughly 18 months. His latest stop is Google's Gemini team, where his expertise in reinforcement learning will be put to work.
Key points
- Barret Zoph joined Google as vice president of research in mid-2026, his fourth senior AI role in roughly 18 months.
- Zoph co-founded Thinking Machines with former OpenAI chief technology officer Mira Murati in late 2024, then left the startup in January 2026 to return to OpenAI.
- He spent only five months at OpenAI before departing in June 2026.
- Google said publicly that it wants Zoph's expertise in reinforcement learning, the training technique that teaches AI models to improve by trial and error, applied to its Gemini AI model.
Barret Zoph has a new job. Again.
The AI researcher has joined Google as vice president of research, the company confirmed to the Wall Street Journal. It is his fourth senior role at a major AI organisation since October 2024.
How did he end up here?
The short version: Zoph left OpenAI in October 2024 after two years there, co-founded a new AI startup called Thinking Machines with Mira Murati (who had been OpenAI's chief technology officer, effectively its top technical officer), then left that startup in January 2026 to go back to OpenAI. He stayed only five months, leading AI sales to large businesses, then left again in June 2026.
Google is not entirely new territory. Zoph worked there before joining OpenAI. "We look forward to Barret returning to Google and bringing his RL and post-training expertise to Gemini," a Google spokesperson said. RL stands for reinforcement learning, a technique where an AI model is rewarded for correct answers and penalised for wrong ones, a bit like training a dog with treats. Post-training refers to the fine-tuning work done on a model after its initial training is complete.
Gemini is Google's flagship AI model, the one powering its AI assistants and competing directly with OpenAI's ChatGPT.
Why does this keep happening across the AI industry?
Executive churn at the top of AI firms is high right now, and OpenAI is the most visible example. The company is preparing for a public stock listing and is one of the most closely watched technology organisations on earth, yet it has lost a string of senior figures over the past eight months, from its chief operating officer to a senior executive overseeing data centres.
No single explanation fits every departure. Pay, creative differences, and the sheer number of well-funded rivals all play a role. When a researcher of Zoph's profile moves, it matters: the knowledge, contacts, and instincts they carry are exactly what companies are spending billions to build.
For ordinary users, the practical question is whether products get better or worse when the people building them move around this fast. The honest answer is that it is too early to say. What is clear is that Google now has a senior researcher who has seen the inside of two of its closest competitors.
Common questions
What is Thinking Machines, the startup Zoph co-founded?
Thinking Machines is an AI startup launched in late 2024 by Mira Murati, the former chief technology officer of OpenAI, and several colleagues including Zoph. It is building AI models and products, though the company has shared few details publicly about its specific focus.
Does Zoph's move affect Google's Gemini AI for regular users?
Not immediately. Research leadership changes take months or years to show up in a finished product. His reinforcement learning background could influence how future versions of Gemini are trained, but any impact on the AI assistant most people use day to day is a long way off.
