People

Author of FlashAttention and co-author of Mamba. Chief scientist at Together AI.

First author of latent diffusion, the method behind Stable Diffusion. Co-founded Black Forest Labs.

Transformer co-author as an intern. Co-founder and CEO of Cohere.

First author of the 2017 paper that trained agents from human preferences, the method behind RLHF. Founded the Alignment Research Center.

First author of the InstructGPT paper, which turned GPT-3 into a model that follows instructions.

Physicist who led the 2020 scaling laws paper. Co-founder and chief science officer of Anthropic.

First author of chain-of-thought prompting and of the emergent abilities paper.

First author of DQN, the network that learned Atari games from pixels.

Led the AlphaGo, AlphaZero and MuZero programs at DeepMind. Reinforcement learning at scale.

Wrote the GPU code and trained AlexNet, which won ImageNet 2012 by a wide margin and started the deep learning era.

Transformer co-author. Later co-founded Adept and Essential AI.

Transformer co-author who pushed for self-attention over recurrence. Co-founded Inceptive.

First author of BERT, the model that made pretrain-then-fine-tune the default for language tasks.

Transformer co-author. Co-founded the NEAR Protocol.

Lead author of the GPT, GPT-2, CLIP and Whisper papers at OpenAI. Few people have shaped more of the current stack.

Co-founder and CEO of Mistral AI. Previously at DeepMind, where he co-authored Chinchilla.

Invented Q-learning in his 1989 thesis. The algorithm behind DQN.

Created Keras and the ARC-AGI benchmark. Argues that scale alone does not give general intelligence.

CTO of OpenAI through ChatGPT and GPT-4. Founded Thinking Machines Lab in 2025.

Co-founder and CEO of Hugging Face, the hub where open models live.

Co-founder and president of OpenAI. Former CTO of Stripe.

Invented generative adversarial networks in 2014 and co-wrote the Deep Learning textbook.

President and co-founder of Anthropic. Earlier at OpenAI and Stripe.

Co-author of AlexNet and seq2seq. Co-founder and chief scientist of OpenAI until 2024, then founder of Safe Superintelligence.

Taught Stanford's first deep learning class, led Autopilot vision at Tesla, founding member of OpenAI. His lectures teach a large share of the field.

CEO of OpenAI. Former president of Y Combinator. Removed and reinstated by the board in November 2023.

Founder of the quant fund High-Flyer and of DeepSeek, whose R1 model in January 2025 showed frontier reasoning at a fraction of the cost.

Led AlphaFold 2, which solved protein structure prediction. Nobel Prize in Chemistry 2024.

First author of Attention Is All You Need, the 2017 paper that introduced the Transformer.

Co-founder of DeepMind and Inflection AI. CEO of Microsoft AI since 2024.

Transformer co-author who named the paper. Co-founded Sakana AI in Tokyo.

Founded Stability AI and funded the public release of Stable Diffusion in 2022.

CEO and co-founder of Anthropic. Led research at OpenAI through GPT-2 and GPT-3. Earlier at Google Brain and Baidu.

Co-author of Stochastic Parrots. Her exit from Google in 2020 became a flashpoint for AI ethics. Founded DAIR.

Go champion who lost four games to one against AlphaGo in March 2016. His game-four win is the last time a human beat a top Go program.

Author of the variational autoencoder and the Adam optimizer, two of the most used tools in the field.

Created word2vec at Google in 2013. Word embeddings with arithmetic: king minus man plus woman equals queen.

Transformer co-author. Later at OpenAI, where he worked on the reasoning models.

Co-founder and CEO of DeepMind. AlphaGo, AlphaFold, Gemini. Nobel Prize in Chemistry 2024.

Created ImageNet, the dataset that made deep learning for vision possible. Co-director of Stanford HAI.

Co-founded Google Brain and Coursera, led Baidu's AI group. His online course introduced millions to machine learning.

Transformer co-author, inventor of the modern mixture-of-experts layer and multi-query attention. Co-founded Character.AI, returned to Google to co-lead Gemini.

Computational linguist. First author of Stochastic Parrots, the standard critique of large language models.

Co-founder and chief AGI scientist of DeepMind. Helped make the term AGI common.

Co-founder and early funder of OpenAI, later its critic and litigant. Founded xAI in 2023.

Probabilistic graphical models at Stanford. Co-founded Coursera and later Insitro.

Google's chief scientist. Co-founded Google Brain, led the work on TPUs, TensorFlow and PaLM, and now Gemini.

Found the vanishing gradient problem in his 1991 thesis and co-invented the LSTM to fix it.

Founder of Mila in Montreal. Neural language models, attention, GANs. Turing Award 2018. Now focused on AI safety.

World chess champion. Lost to IBM's Deep Blue in 1997, the first defeat of a reigning champion by a machine under match conditions.

Co-founder and CEO of Nvidia. Bet the company on GPU computing, then on AI.

Co-inventor of the LSTM. Long-time director at IDSIA. Argues that much of deep learning was invented earlier than credited.

Co-author of the standard AI textbook with Peter Norvig. Leading voice on the control problem.

Co-author of the 1995 support-vector networks paper. Head of Google Research in New York.

Built convolutional networks that read handwritten digits at Bell Labs (LeNet). Chief AI Scientist at Meta. Turing Award 2018.

Machine learning statistician at Berkeley. Variational methods, graphical models, and the advisor of a generation of researchers.

Father of modern reinforcement learning: temporal-difference learning, the standard textbook with Andrew Barto, and the essay The Bitter Lesson. Turing Award 2024.

Co-author with Richard Sutton of Reinforcement Learning: An Introduction. Turing Award 2024.

The most cited figure in deep learning. Backpropagation (1986), Boltzmann machines, deep belief nets (2006), AlexNet (2012). Turing Award 2018, Nobel Prize in Physics 2024.

Described backpropagation for training neural networks in his 1974 Harvard thesis, twelve years before it caught on.

Computational neuroscientist. Co-inventor of the Boltzmann machine with Hinton and Ackley.

Cognitive scientist. Lead author of the 1986 backpropagation paper and the Parallel Distributed Processing volumes.

Co-inventor of the support vector machine and of statistical learning theory (VC dimension).

Built the theory of Bayesian networks and causal inference. Turing Award 2011.

Father of expert systems. DENDRAL and MYCIN showed that encoded expert knowledge could solve narrow problems.

Designed the Neocognitron in 1980, the first convolutional network, inspired by the visual cortex.

Physicist whose 1982 network stored memories as energy minima. Shared the 2024 Nobel Prize in Physics with Geoffrey Hinton.

Built ADALINE in 1960 with Ted Hoff and gave neural networks the least-mean-squares learning rule.

Psychologist who built the perceptron in 1958, the first machine that learned to classify from examples.

Mathematician and educator. Co-author of Perceptrons and creator of the Logo language.

Statistician who invented random forests and bagging, and who told statisticians to take algorithmic models seriously.

Coined the term artificial intelligence for the 1956 Dartmouth workshop. Invented Lisp and founded the Stanford AI Lab.

Co-founder of the MIT AI Lab. His 1969 book Perceptrons, with Seymour Papert, showed the limits of single-layer networks and cooled interest in them for a decade.

Co-author of the Logic Theorist and the General Problem Solver. A founder of symbolic AI and cognitive science.

Applied mathematician whose 1973 report for the UK government judged AI research a failure and cut its funding. It started the first AI winter.

Self-taught logician. Co-author of the 1943 McCulloch-Pitts paper at age 20.

Wrote ELIZA in 1966, the first chatbot. Later a critic of the field he helped start.

Economist and computer scientist. With Allen Newell he wrote the Logic Theorist in 1956, the first program that proved theorems. Nobel in economics, Turing Award.

Founder of information theory. His 1948 paper defined the bit and entropy, the measure every language model is trained to reduce.

Ukrainian scientist who trained deep multilayer networks in the 1960s with the Group Method of Data Handling.

Mathematician who defined computation and, in 1950, asked whether machines can think. He proposed the imitation game, now called the Turing test.

Mathematician behind the stored-program computer architecture and game theory. His last book compared the computer and the brain.

IBM engineer whose checkers program learned from play. He used the phrase machine learning in 1959.

Co-author of Mamba and the structured state space models that led to it.

Neurophysiologist who, with Walter Pitts, wrote the 1943 paper that described neurons as logical units. The artificial neuron starts here.

Mathematician who founded cybernetics in 1948, the study of control and communication in animals and machines.

Stanford professor behind SQuAD, HELM and the Center for Research on Foundation Models, which named the category.

First author of Denoising Diffusion Probabilistic Models, the 2020 paper that made diffusion models work.

Lead author of ResNet, which made networks hundreds of layers deep trainable. Also Mask R-CNN and masked autoencoders.