Barret Zoph, who previously abruptly left and joined OpenAI due to a disagreement with Thinking Machines Lab CEO Miriam Mularati, has once again changed employers. Alphabet's Google announced that Zoph will return as Vice President of Research, bringing his expertise in reinforcement learning and post-training into the Gemini team.

The cycle of a "revolving door" researcher

Zoph's career is a typical example of AI talent mobility. He joined Google Brain in 2016 for large model research, moved to OpenAI in 2022 to participate in the development of ChatGPT and led the post-training and foundational model productization, became CTO of Thinking Machines Lab in 2024, and returned to OpenAI in January 2026.

In January of this year, Luke Metz and Sam Altman, co-founders of Thinking Machines Lab, also left the company, and all three subsequently joined OpenAI — and before starting their own business, they had all worked at OpenAI. A Google spokesperson said they were looking forward to his return with experience in RL and post-training, which also reflects Gemini's urgent need for top post-training talent during its sprint phase, and how researchers' affiliations have become a key factor in the competition for large models.