Article
AI News

Barret Zoph joins DeepMind as vice president of research

The co‑founder of Thinking Machines Lab and a former vice president of research at OpenAI will join Alphabet’s DeepMind as vice president of research.

by Whatsnew Newsroom
The image features a close-up view of computer code with a prominent logo of Google DeepMind. The background consists of dark, colorful lines of code that suggest a focus on artificial intelligence and machine learning development.

Another senior AI researcher has moved between the handful of organisations that now set the field’s agenda, underscoring how talent continues to circulate between startups and the big labs.

Barret Zoph will join Google's DeepMind as vice president of research, the company said. He previously served as Vice President of Research at OpenAI and, according to his LinkedIn profile, returned to OpenAI briefly after he left Thinking Machines earlier this year.

Zoph began his career in Google research, having spent six years with Google as a research scientist in Google Brain before that group was merged into DeepMind. His hiring is the latest move amid an Alphabet reshuffle of its AI teams: this month DeepMind CEO Dennis Hassabis was reported as stepping aside, and the company disclosed that teams were moving out of DeepMind and into corporate Google.

For users and customers the practical effect of a single hire is limited; for the industry it is another data point in a contest for researchers who can lead large-scale model and systems work. At DeepMind the appointment signals a reinforcement of research leadership as the lab’s structure and reporting lines are rearranged inside Alphabet.

What to watch next: whether Zoph’s remit will fold teams or projects back under DeepMind’s research umbrella, and how DeepMind’s leaders slot into the broader corporate AI organisation as the reorganisation continues.

by Whatsnew Newsroom
whatsnew. APPS · WEB TOOLS · SECURITY · AI

Know what’s new.

The useful side of the internet. Covered properly.

Set as preferred →