OpenAI co-founder Andrej Karpathy joins Anthropic

Andrej Karpathy started this week at Anthropic, where he is working on pre-training under team lead Nick Joseph. Pre-training is the large-scale process that gives Anthropic’s Claude models their core knowledge and abilities. It is also one of the most expensive and computing-intensive parts of building advanced AI models.

“I’ve joined Anthropic,” Karpathy posted on X Tuesday. “I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D.”

Andrej Karpathy is widely known for combining deep AI research knowledge with practical experience in training large-scale AI models. His hiring suggests that Anthropic believes AI-assisted research and innovation will help it compete with companies like OpenAI and Google.

At OpenAI, Karpathy worked on deep learning and computer vision before leaving in 2017 to join Tesla, where he led the companys Autopilot and Full Self-Driving programs. He left Tesla in 2022, returned to OpenAI for a year, and then departed again in 2024 to launch Eureka Labs, a startup focused on using AI assistants in education.

Andrej Karpathy has shared only limited updates about Eureka Labs since launching it, and it is unclear whether he will remain involved with the startup. He also teaches an online course called Neural Networks: Zero to Hero, which helps students learn to build neural networks from scratch, and regularly shares lectures on LLMs and AI on his YouTube channel.

Andrej Karpathy was one of several OpenAI employees who later worked at Tesla to help improve the company’s self-driving technology, which was developing more slowly than expected. Karpathy left Tesla in 2022, and the company still does not sell a vehicle that can safely operate without a human driver ready to take control at any time.

Read More:Relay Raises $50M in Funding Led by General Catalyst

Separately, Anthropic has hired Chris Rohlf to join its frontier red team, which tests advanced AI models against serious security threats. Rohlf has more than 20 years of cybersecurity experience and previously worked at Yahoo on its cybersecurity team known as “The Paranoids.” He later spent six years at Meta before joining Anthropic. Rohlf was also a fellow at Center for Security and Emerging Technology at Georgetown University, where he worked on the CyberAI project.

“We have a real opportunity in front of us to dramatically improve cyber security with AI,” Rohlf said in a post on X. “I can’t think of a better company or team to join at this critical moment in time.”

Read More:Brami Raises $33M in Series B Funding Led by VMG Partners