François Chollet (born October 20, 1989) is a French software engineer, computer scientist, and artificial intelligence researcher active from 2012 to the present, who was educated at ENSTA Paris (ENSTA ParisTech) where he earned a Master of Engineering degree.^[13,33,2,1] While multiple sources state that he is a former Senior Staff Engineer at Google who worked there from 2015 to 2024, the Albanian edition reports that he is currently working at Google.^[33,2,1]

Chollet is best known as the creator and primary author of Keras, a deep learning program library in the Python programming language released in 2015 that facilitates interaction and rapid experimentation with neural networks and is integrated into Google's TensorFlow library.^[14,33,2,1] His other significant projects and contributions include major contributions to the TensorFlow machine learning framework, developing the XCeption neural network model, initiating the ARC Prize (or ARC Challenge), and authoring several books on deep learning.^[33] His research focuses on computer vision, the practical application of machine learning to formal reasoning and abstraction, and achieving greater generality in artificial general intelligence (AGI or strong artificial intelligence).^[2,1,13] He has received notable recognitions including the Global Swiss AI Award and inclusion in the TIME100 AI list.

In 2012, Chollet graduated with an engineering degree (Diplôme d'Ingénieur / Master of Engineering) from ENSTA Paris (École nationale supérieure de techniques avancées), an institution affiliated with the Institut Polytechnique de Paris.^[15,25] In 2012, he was active as a researcher in computer vision at the University of Tokyo, where he participated with his team in the ImageNet competition won by Geoffrey Hinton's students Alex Krizhevsky and Ilya Sutskever, developing his interest in deep learning.^[35]

Starting in August 2015, shortly after releasing the open-source deep learning library Keras, which allows rapid experimentation with neural networks, Chollet began working for Google LLC in Mountain View, California.^[25] He is the author of several books on deep learning, including Deep Learning with Python, and his research papers have been published in major conferences such as CVPR, NeurIPS, and ICLR.^[25,7] Chollet also developed the Xception neural network model and authored the paper "Xception: Deep Learning with Depthwise Separable Convolutions", which ranks among the top ten most cited papers in the CVPR proceedings.^[17,8,3,9]

In 2019, he formulated and published the Abstraction and Reasoning Corpus for Artificial General Intelligence (ARC-AGI) benchmark, an AI test featuring a set of visual problems designed to measure the ability of AI systems to solve novel reasoning tasks.^[3,17,8,26] In this context, Chollet defined the intelligence of a technical system as a measure of its skill-acquisition efficiency over a scope of tasks, with respect to priors, experience, and generalization difficulty. In an interview expressing his assessment of artificial intelligence, he took the view that investments in generative AI are exaggerated and based on false promises.^[19,18] In 2024, Chollet launched the ARC Prize, a US$1 million competition to solve the ARC-AGI benchmark.^[4,5] In early 2025, he announced the expansion of the ARC Prize into a full-fledged non-profit foundation to guide and accelerate research progress toward artificial general intelligence.^[27]

In November 2024, after spending more than nine years at Google, Chollet left the company to pursue his own projects in the field of artificial general intelligence.^[5,6,4] Along with Zapier co-founder Mike Knoop, he co-founded a new startup focused on developing artificial general intelligence using program synthesis.^[5,6,4]

In 2019, Chollet published the ARC-AGI benchmark for AI systems, and in 2024, he initiated the creation of the ARC Prize, a competition for solving the ARC-AGI benchmark with a million-dollar prize fund.^[26,4] In the same year, he left Google after working there for nine years.^[5] In early 2025, Chollet announced the transformation of the ARC Prize into a full-fledged non-profit foundation that will engage in research in the field of artificial general intelligence.^[27]

François Chollet's artificial intelligence research papers have been published at major flagship conferences in the field, including the Conference on Computer Vision and Pattern Recognition (CVPR), the Conference on Neural Information Processing Systems (NeurIPS), and the International Conference on Learning Representations (ICLR).^[7] He is the author of the paper "Xception: Deep Learning with Depthwise Separable Convolutions", which ranks among the top ten most cited research papers in CVPR proceedings with over 18,000 citations.^[9,8] He is also the creator and developer of the Abstraction and Reasoning Corpus (ARC) Challenge.^[26,10]

Chollet is the author of the book Deep Learning with Python, which has sold over 100,000 copies.^{[10,11,12,26]} He also co-authored the book Deep Learning with R, with sources identifying his co-author as either Tomasz Kalinowski or Joseph J. Allaire.^{[10,11,12,26]}

On December 1, 2021, François Chollet won the Global Swiss AI Award (also referred to as the Swiss Global AI Award) for outstanding achievements, advances, and breakthroughs in the field of artificial intelligence technology.^[2,1,28] In September 2024, TIME magazine named and ranked Chollet as one of the 100 most influential people in artificial intelligence.^[2,1,29]

François Chollet's publications include Deep Learning with Python, published by Manning Publications in 2017, and its second edition published in 2021. Together with J. J. Allaire, he authored Deep Learning With R, which was published by Manning Publications in 2018. In 2019, Chollet authored On the Measure of Intelligence.