- Andrej Karpathy, OpenAI co-founder and former Tesla AI lead, has joined Anthropic to work on pre-training under team lead Nick Joseph
- Anthropic says Karpathy will start a new team focused on using Claude to accelerate pre-training research — the most expensive phase of building frontier models
- Cybersecurity veteran Chris Rohlf, formerly of Yahoo’s red team and Meta, has also joined Anthropic’s frontier red team
Andrej Karpathy, the AI researcher who co-founded OpenAI and previously led AI at Tesla, has joined Anthropic, per his personal page. He announced the move on X Tuesday, writing that the next few years at the frontier of LLMs will be especially formative and that he is excited to get back to R&D.
Karpathy started this week at Anthropic, where he is working on pre-training under team lead Nick Joseph. An Anthropic spokesperson told TechCrunch that Karpathy will start a team focused on using Claude to accelerate pre-training research. Pre-training is the large-scale training run that gives a model its core knowledge — and it is by far the most expensive, compute-intensive phase of building a frontier model.
The move is Karpathy’s third major career pivot in three years. He returned to OpenAI in February 2023 to build a team working on midtraining and synthetic data generation, then left again in February 2024 to start Eureka Labs, an AI-native education startup. He has not shared many updates on Eureka Labs since its launch, and said Tuesday he remains deeply passionate about education and plans to resume his work on it in time.
Separately, Anthropic has also brought on Chris Rohlf to its frontier red team, which stress-tests advanced AI models against severe threats. Rohlf spent six years at Meta and previously led Yahoo’s well-respected cybersecurity team known as The Paranoids. He was also a fellow at Georgetown’s Center for Security and Emerging Technology. Rohlf said he sees a real opportunity to dramatically improve cybersecurity with AI.
The hires signal that Anthropic is pushing on two fronts simultaneously: accelerating the compute-heavy training that keeps it competitive with OpenAI and Google, and hardening its models against adversarial attacks. Bringing in Karpathy for the first and Rohlf for the second is a clear statement of intent.
[Editor’s note: This article was updated on May 19, 2026 to correct the description of Karpathy’s career pivot timeframe. The original text stated “in two years”; corrected to “in three years” to accurately reflect the period from his return to OpenAI in February 2023 through his joining Anthropic in May 2026.]
FAQ
What will Karpathy do at Anthropic?
He will work on pre-training under team lead Nick Joseph and start a new team focused on using Claude to accelerate pre-training research — essentially using AI to make AI training faster and more efficient.
What happened to Eureka Labs?
Karpathy founded Eureka Labs in 2024 as an AI-native education startup, but has not shared many updates since its launch. He said Tuesday he remains passionate about education and plans to resume his work on it in time.
Who is Chris Rohlf?
A cybersecurity veteran with 20+ years of experience. He previously led Yahoo’s red team (The Paranoids), spent six years at Meta, and was a fellow at Georgetown’s Center for Security and Emerging Technology. He joins Anthropic’s frontier red team.
