Andrej Karpathy Joins Anthropic to Lead AI-Assisted Pre-Training Research
Andrej Karpathy Joins Anthropic for AI Pre-Training

Andrej Karpathy, one of the original 11 co-founders of OpenAI and former head of Tesla's AI division, has officially joined Anthropic. He announced the move in a post on X, sending ripples through the AI industry. Karpathy started this week on Anthropic's pre-training team, which handles the large-scale training runs responsible for giving Claude its core knowledge and capabilities.

New Role and Responsibilities

Karpathy will build and lead a new group focused on using Claude itself to accelerate pre-training research. This role sits squarely at the intersection of AI-assisted research and frontier model development. He will report to Nick Joseph, Anthropic's head of pre-training and a former OpenAI alumnus.

A Stellar Career in AI

The 39-year-old Slovak-Canadian researcher earned his PhD at Stanford under Fei-Fei Li and co-designed the university's first deep learning course. At Tesla, from 2017 to 2022, he led the computer vision team behind Autopilot, overseeing everything from in-house data labeling to neural network deployment on Tesla's custom inference chip. He returned to OpenAI from 2023 to 2024, where he built a team focused on midtraining and synthetic data generation, experience that maps directly onto his new role.

Wide Pickt banner — collaborative shopping lists app for Telegram, phone mockup with grocery list

Since leaving OpenAI last year, Karpathy has been one of AI's most visible public educators. He runs a popular YouTube channel, posts technical explainers, and launched Eureka Labs, an AI-native school whose first product walks students through training their own AI system. He also coined the term "vibe coding."

Implications for Anthropic

The hire signals that Anthropic believes the path to staying competitive with OpenAI and Google runs through AI-assisted research, not just raw compute. Karpathy is one of the rare figures who can bridge theoretical LLM research with the messy practicalities of large-scale training.

"I think the next few years at the frontier of LLMs will be especially formative," Karpathy wrote. As for Eureka Labs and his education work, he says he remains passionate and plans to return to it, but not yet.

Pickt after-article banner — collaborative shopping lists app with family illustration