Andrej Karpathy

Andrej Karpathy is a Slovak-Canadian computer scientist, and AI researcher, and was also a key member of the founding team at OpenAI, Director of AI at Tesla.

Andrej is currently focused on AI education through his company, Eureka Labs. Founded in 2024, Eureka Labs is dedicated to teaching AI concepts, with a focus on large language models (LLMs). Before this, Karpathy was closely involved with OpenAI, where he played a key role in building a team and improving GPT-4 for ChatGPT. He returned to OpenAI in 2023 after an initial stint from 2015 to 2017 as a founding member and research scientist, before leaving again in 2024 to concentrate on his projects.

Karpathy is widely recognised for his work at Tesla, where he served as the Senior Director of AI from 2017 to 2022. He led the computer vision team responsible for Tesla’s Autopilot system, including in-house data labelling, neural network training, and deploying the system to Tesla's fleet. His team’s efforts focused on developing Full Self-Driving capabilities, significantly improving driver assistance and safety. Notable events where he presented this work include Tesla AI Day in 2021 and Tesla Autonomy Day in 2019, providing insights into the company’s progress in AI-driven autonomous systems.

In addition to his work in industry, Karpathy has made notable contributions to AI research and education. He is the author and primary instructor of Stanford University’s first deep learning course, CS 231n: Convolutional Neural Networks for Visual Recognition. This course, which started in 2015, rapidly grew in popularity, becoming one of the largest classes at Stanford, with hundreds of students enrolling annually.

Karpathy’s research has produced influential publications in the field of deep learning. His works include research on deep reinforcement learning, convolutional and recurrent neural networks, and image captioning systems. Some of his most cited papers include "DenseCap: Fully Convolutional Localization Networks for Dense Captioning", "PixelCNN++: A PixelCNN Implementation", and his PhD thesis, "Connecting Images and Natural Language". His work on large-scale video classification using convolutional neural networks also gained significant attention in the AI community.

Karpathy completed his PhD at Stanford University in 2015, under the supervision of Fei-Fei Li at the Stanford Vision Lab. His research focused on the intersection of computer vision and natural language processing. During his academic career, he also worked with prominent researchers such as Daphne Koller, Andrew Ng, and Sebastian Thrun. Karpathy holds a master’s degree from the University of British Columbia, where he worked on machine learning for agile robotics, and a bachelor’s degree in computer science and physics from the University of Toronto.

Andrej Karpathy was born on 23 October 1986 in Bratislava, Czechoslovakia, which is now Slovakia. At the age of 15, he moved with his family to Toronto, Canada, where he completed his education. Karpathy developed an early interest in computer science and physics, which led him to pursue a Bachelor of Science degree at the University of Toronto. During his undergraduate studies, he majored in computer science and physics and took part in Geoff Hinton’s deep learning classes. This exposure introduced him to the emerging field of artificial intelligence, specifically deep learning, which shaped his future career.

After completing his undergraduate degree in 2009, Karpathy pursued a Master of Science at the University of British Columbia. His research focused on machine learning for agile robotics, where he worked on learning controllers for physically simulated figures under the guidance of Michiel van de Panne. His master's work revolved around creating models for simulated movement, which laid the foundation for his understanding of machine learning applications in physical environments.

In 2011, Karpathy started his PhD at Stanford University under the supervision of Fei-Fei Li. His research centred around the intersection of computer vision and natural language processing. His work involved building neural networks that could describe images with natural language, a critical task in AI. During this time, he worked with several notable researchers, including Daphne Koller, Andrew Ng, and Sebastian Thrun. In 2016, he completed his PhD with a thesis titled "Connecting Images and Natural Language", which remains an influential piece of work in the field of AI. His time at Stanford also saw him teach the first deep learning course at the university, CS 231n: Convolutional Neural Networks for Visual Recognition. The course quickly gained popularity and became one of the largest at Stanford.

In 2015, Karpathy was a founding member of OpenAI, an AI research organisation created to ensure artificial general intelligence benefits humanity. His research at OpenAI focused on deep learning models, especially in the areas of reinforcement learning, natural language processing, and computer vision. He contributed to several papers, including "DenseCap: Fully Convolutional Localization Networks for Dense Captioning" and "PixelCNN++", which gained recognition within the AI community.

In June 2017, Karpathy joined Tesla as Senior Director of AI. At Tesla, he led the computer vision team responsible for developing the Autopilot system. His role involved managing in-house data labelling, neural network training, and deploying these networks into production. The team worked on Full Self-Driving (FSD) technology, which aimed to enable autonomous driving for Tesla vehicles. Karpathy’s work contributed to several advancements in Tesla's AI-driven systems, improving the safety and convenience of its cars. He presented these efforts at several key events, including Tesla Autonomy Day in 2019 and Tesla AI Day in 2021, where he provided a detailed overview of Tesla’s AI progress.

During his time at Tesla, Karpathy also contributed to the development of AI-related projects and technologies within the company. His work at Tesla was instrumental in improving their AI systems, particularly in advancing Full Self-Driving capabilities.

After a sabbatical in 2022, Karpathy decided to leave Tesla in July of that year, choosing to focus on personal projects. In 2023, he returned to OpenAI, where he contributed to the improvement of GPT-4 and its integration into ChatGPT. During this period, he built a small team and focused on enhancing the model’s performance and capabilities.

In 2024, Karpathy left OpenAI to start his own AI-focused education company called Eureka Labs. The goal of Eureka Labs is to teach AI and large language models (LLMs) to students, focusing on making advanced AI concepts more accessible and understandable. Karpathy’s company is aimed at bridging the gap between AI research and education, bringing his years of experience in AI development into the educational sector.

Andrej Karpathy’s vision focuses on making artificial intelligence accessible to everyone through education and practical applications. He aims to bridge the gap between advanced AI technologies and students or researchers by simplifying complex concepts. With his new venture, Eureka Labs, Karpathy seeks to empower the next generation of AI developers by providing them with the tools and knowledge needed to innovate in the field. His vision also extends to ensuring that AI is used responsibly, benefitting society while encouraging transparency, collaboration, and ethical development in the field of artificial intelligence.

Andrej Karpathy has received several notable awards and recognitions for his contributions to AI and deep learning. In 2020, he was named one of MIT Technology Review's Innovators Under 35, recognising his work in advancing AI technologies. He was also featured in Time Magazine's 2024 list of the 100 Most Influential People in AI. Karpathy's work in Tesla’s Autopilot system and his role as a founding member of OpenAI have been widely acknowledged in the tech community. His teaching at Stanford University and contributions to AI education have further solidified his influence in the field.