Andrej Karpathy

Andrej Karpathy is one of the most influential educators and practitioners in modern deep learning. After completing his PhD with Fei-Fei Li at Stanford, he helped shape early research culture at OpenAI and later led the computer vision team for Tesla Autopilot, where he bridged cutting-edge research and large-scale deployed systems. He is known for unusually clear explanations of neural networks, training dynamics, and how “software 2.0” changes engineering practice.

His public writing and talks connect first principles—backpropagation, sequence models, reinforcement learning from human feedback—with practical intuition for how today’s large models are built and debugged. Whether he is unpacking tokenization, discussing data flywheels for autonomous driving, or narrating the mental model behind a transformer, the through-line is rigorous thinking presented in plain language.

Karpathy is worth following if you want a practitioner’s map of the field: not only what works in benchmarks, but how teams iterate, what breaks in production, and how research ideas translate into products. His mix of research depth, systems experience, and teaching skill makes his content especially valuable for engineers and researchers who care about both theory and shipping.