Andrej Karpathy

There are people who build AI, and there are people who truly explain AI. Andrej Karpathy is one of the rare few who do both simultaneously, and at the same level of depth. Born in Bratislava in 1986, he moved to Toronto as a teenager and went on to complete his PhD at Stanford under Fei-Fei Li, the same researcher who transformed computer vision with ImageNet. From that point on, his trajectory never slowed down.

In 2015 he was one of the co-founders of OpenAI, the laboratory that later launched ChatGPT to the world. After two years there as a research scientist, he took on a different challenge: joining Tesla as Director of AI to build the Autopilot system, the one that today lets cars recognize signs, pedestrians, and lanes in real time. He achieved something very few people can claim: major impact in both academia and in a product that millions of people use every day. In 2022 he left Tesla, briefly returned to OpenAI, then founded Eureka Labs in July 2024, an AI education platform whose flagship course, LLM101n, teaches you to build your own language model from scratch using Python and C. In May 2026 he joined Anthropic to lead pretraining research, meaning he is working directly on the core of what makes Claude so capable.

But if there is one thing that defines Karpathy more than any title or company, it is his ability to teach. His CS231n course at Stanford on convolutional neural networks for visual recognition was the first deep learning course offered at Stanford, growing from 150 enrolled students in 2015 to 750 in 2017. His YouTube tutorials, including "Neural Networks: Zero to Hero," have been watched millions of times and are considered by many to be the best hands-on introduction to the subject that exists. His blog posts, articles like "The Unreasonable Effectiveness of Recurrent Neural Networks" (2015), and projects like nanoGPT, a minimalist GPT in fewer than 300 lines of code, are the resources that open doors for anyone who genuinely wants to understand how all of this works.

The lesson Karpathy teaches me goes beyond any technique: the most valuable talent is not doing incredible things in silence, but doing incredible things and then explaining them so clearly that others can carry the work forward. Every time someone takes their first neural networks class and thinks "wait, I actually get this," there is a very good chance Andrej Karpathy had something to do with it. And those same students, using tools like Claude, are building the next generation of what comes next.