Andrej Karpathy Joins Anthropic to Boost AI Pretraining Research

The high-stakes world of artificial intelligence saw a major talent migration this week. Andrej Karpathy, a renowned Slovak-Canadian AI researcher, announced on Tuesday, May 19, 2026, that he has officially joined Anthropic, a key player in the large language model (LLM) arena.

This strategic move places Karpathy on Anthropic's pretraining team, where he will operate under the leadership of Nick Joseph, Anthropic's Head of Pretraining and a former OpenAI alumnus. His primary role involves spearheading a new team dedicated to utilizing Anthropic's flagship AI model, Claude, to accelerate pretraining research. The ambitious goal is to push the boundaries of recursive self-improvement within AI systems, a critical step toward more advanced artificial intelligence.

Karpathy's return to the AI research frontier follows a diverse career path that has seen him contribute significantly to both industry titans and educational initiatives. Born on October 23, 1986, in Bratislava, Czechoslovakia, he immigrated to Toronto at the age of 15. His academic journey includes bachelor's degrees in Computer Science and Physics from the University of Toronto in 2009, a master's degree from the University of British Columbia in 2011, and a PhD from Stanford University in 2015. At Stanford, under the tutelage of Fei-Fei Li, he specialized in the intricate intersection of natural language processing, computer vision, and deep learning models. Notably, he authored and served as the primary instructor for Stanford's first deep learning course, CS 231n, which quickly became one of the university's largest and most influential classes.

His professional resume boasts impactful stints at some of the most influential organizations in artificial intelligence. From 2015 to 2017, Karpathy was a founding member and research scientist at OpenAI, where his work spanned deep learning, computer vision, generative modeling, and reinforcement learning. In June 2017, he transitioned to Tesla, taking on the role of Director of Artificial Intelligence and Autopilot Vision, reporting directly to Elon Musk. There, he led the computer vision team responsible for Tesla's Autopilot system and briefly for Tesla Optimus, overseeing in-house data labeling, neural network training, and deployment on custom inference chips. After a sabbatical, he departed Tesla in July 2022.

Karpathy later returned to OpenAI in February 2023, where he established a new team focused on midtraining and synthetic data generation, before again leaving in February 2024. In July 2024, he founded Eureka Labs, an AI education platform that launched its inaugural product, the AI course LLM101n. Beyond his corporate roles, Karpathy has been a prominent AI educator through his popular YouTube videos and the