Fei-Fei Li

ref · May 25, 2026, 6:01am

Inaugural Sequoia Professor of Computer Science at Stanford University, creator of ImageNet, and founding co-director of the Stanford Institute for Human-Centered AI — widely recognized as one of the central architects of the modern deep learning revolution.

Profile

Field	Detail
Born	July 3, 1976, Beijing, China
Nationality	Chinese-American
Current Institutions	Stanford University (on partial leave); World Labs (Co-founder & CEO)
Research Areas	Computer Vision, Deep Learning, AI in Healthcare, Cognitive Neuroscience, Spatial Intelligence, Robotic Learning
PhD Advisors	Pietro Perona (primary); Christof Koch (secondary)
PhD Dissertation	Visual Recognition: Computational Models and Human Psychophysics (Caltech, 2005)
Stanford Profile	profiles.stanford.edu/fei-fei-li
X / Twitter	@drfeifei
Google Scholar	scholar.google.com

Overview

Fei-Fei Li (李飞飞) is a Chinese-American computer scientist whose creation of ImageNet — a large-scale annotated visual dataset encompassing over 14 million images — is widely credited as one of the three foundational catalysts of the modern AI and deep learning revolution. She is the inaugural Sequoia Professor in the Computer Science Department at Stanford University and a founding co-director of the Stanford Human-Centered Artificial Intelligence Institute (HAI). Beyond her research contributions, she has shaped an academic generation through her doctoral mentorship of figures including Andrej Karpathy, Timnit Gebru, and Olga Russakovsky, and through her widely-adopted CS231n course on deep learning for computer vision. In 2024 she co-founded World Labs, a startup focused on spatial intelligence that reached a valuation exceeding $1 billion within months of founding and raised a further $1 billion in 2026. She has been described informally as the “godmother of AI,” a designation she accepts as recognition of women’s contributions to the field.

Early Life & Education

Li was born in Beijing in 1976 and grew up in Chengdu, Sichuan, attending Sichuan Chengdu No. 7 High School. When she was twelve, her father emigrated to Parsippany, New Jersey; at sixteen, Li and her mother joined him. At Parsippany High School, she worked weekends at the family dry-cleaning business. A high school mathematics teacher who recognized her ability proved pivotal, helping her pursue university admission in an unfamiliar system. She was later inducted into Parsippany High School’s hall of fame in 2017.

B.A., Physics — Princeton University, 1999
Li majored in physics at Princeton, completing a senior thesis on computational auditory modeling under Bradley Dickinson. During her undergraduate years she returned home most weekends to help run the family dry-cleaning business and worked as a dishwasher to supplement income, while holding a Paul & Daisy Soros Fellowship for New Americans.

M.S., Electrical Engineering — California Institute of Technology (Caltech), 2001
Graduate study at Caltech brought Li under the supervision of Pietro Perona, whose work on visual recognition shaped the direction of her research.

Ph.D., Electrical Engineering — California Institute of Technology (Caltech), 2005
Her dissertation, Visual Recognition: Computational Models and Human Psychophysics, was supervised primarily by Pietro Perona, with secondary supervision from neuroscientist Christof Koch. The work bridged machine and human approaches to rapid scene understanding — a methodological stance that would define her career.

Career

University of Illinois Urbana-Champaign (2005–2006)

Li joined UIUC’s Electrical and Computer Engineering Department as an assistant professor immediately after completing her doctorate.

Princeton University (2007–2009)

Returning to Princeton as an assistant professor in Computer Science, Li began developing what would become ImageNet. Inspired by cognitive psychologist Irving Biederman’s estimate that humans recognize approximately 30,000 object categories, she conceived a dataset of comparable scale. The project initially met skepticism from peers who considered labeling millions of images impractical; Li persisted, eventually employing Amazon Mechanical Turk to annotate the corpus.

Stanford University (2009–present)

Li joined Stanford as an assistant professor in 2009, was promoted to associate professor with tenure in 2012, and to full professor in 2018. She holds the inaugural Sequoia Professorship in Computer Science.

Stanford Artificial Intelligence Lab — Director (2013–2018)
Li served as director of SAIL during a period of rapid expansion in the field, overseeing research and representing the lab publicly.

Stanford HAI — Co-Founding Director (2019–present)
Li co-founded Stanford’s Human-Centered AI Institute alongside former provost John Etchemendy, establishing a university-wide initiative linking AI research with questions of policy, ethics, education, and human benefit. The institute has become one of the most prominent centers for AI governance research in the United States.

ImageNet Large Scale Visual Recognition Challenge — Organizer (2010–2017)
Li led the team organizing ILSVRC (ImageNet Large-Scale Visual Recognition Challenge), the annual competition that served as the proving ground for AlexNet in 2012 and catalyzed the deep learning era in computer vision.

CS231n — Course creator
Li created Stanford’s CS231n, “Deep Learning for Computer Vision,” which became one of the most widely-adopted AI courses globally and has been viewed and studied by millions of practitioners.

RAISE-Health — Co-Launch (2023)
Li co-launched the Responsible AI for Safe and Equitable Health (RAISE-Health) initiative at Stanford in collaboration with Stanford Medicine, developing frameworks for responsible AI deployment in clinical care and biomedical research.

Google Cloud (January 2017–September 2018)

On sabbatical from Stanford, Li joined Google Cloud as Vice President and Chief Scientist of AI/ML. Her team worked to lower the barrier for businesses and developers to adopt AI, including contributing to the AutoML product line. The period was complicated by Google’s involvement in Project Maven, a Pentagon drone-imaging contract; Li privately expressed concern about public perception of the military AI connection, and later publicly reaffirmed her commitment to human-centered AI principles.

AI4ALL — Co-Founder (2017–present)

Li co-founded AI4ALL, a nonprofit dedicated to broadening participation in AI, with Melinda French Gates and Jensen Huang. The organization grew from a precursor Stanford summer program (SAILORS) focused on ninth-grade girls and expanded to programs at Princeton, Carnegie Mellon, Boston University, UC Berkeley, and Simon Fraser University. Li serves as chairperson.

World Labs — Co-Founder and CEO (2024–present)

In 2024, Li co-founded World Labs with three colleagues to build AI systems for “spatial intelligence” — the capacity to reason about and act within three-dimensional physical environments, integrating visual perception with real-world interaction. The company raised $230 million in seed funding in September 2024, reached a valuation above $1 billion within four months of its founding, and raised an additional $1 billion in February 2026.

Key Contributions

ImageNet — Conceived and built starting in 2007, ImageNet provided over 14 million labeled images across 22,000 categories, and the ILSVRC competition it spawned served as the direct catalyst for AlexNet’s breakthrough in 2012 and the subsequent deep learning revolution in computer vision, autonomous vehicles, medical imaging, and language-vision models. The CVPR paper introducing ImageNet (Deng, Dong, Socher, Li-Jia Li, Kai Li, Fei-Fei Li, 2009) is among the most cited in the history of computer science.
ImageNet Large Scale Visual Recognition Challenge (ILSVRC) — Li organized the annual competition from 2010 to 2017; the 2012 edition, in which AlexNet achieved a top-5 error rate roughly halving the prior best, is widely regarded as the moment that launched the modern deep learning era.
CS231n: Deep Learning for Computer Vision — Li’s Stanford course became an industry-standard curriculum for applied computer vision, training a generation of practitioners and researchers; the 2015 version attracted millions of online viewers.
Human-Centered AI (HAI) framework — Li articulated and institutionalized the concept of human-centered AI — advancing research, education, policy, and practice with explicit attention to human benefit, equity, and societal impact — through Stanford HAI and through extensive public advocacy including U.S. Congressional testimony (2018) and the opening keynote of the Paris AI Action Summit (2025).
Spatial Intelligence research agenda — Through her academic lab and World Labs, Li has advanced spatial intelligence as a distinct research frontier: enabling AI systems to perceive, model, and reason within three-dimensional physical environments, with applications in robotics, embodied agents, and generative world models.
AI in Healthcare — Li’s collaborations with Stanford Medicine and the RAISE-Health initiative have produced frameworks for clinical AI deployment, patient safety, and equitable health AI, influencing how academic medicine approaches responsible adoption of AI systems.
AI4ALL — The nonprofit co-founded by Li has run diversity-focused AI education programs at major universities, directly reaching thousands of high school students from underrepresented groups and producing a cohort of AI practitioners who would not otherwise have entered the field.
One-Shot Learning and Scene Understanding — Li’s early academic work on learning object categories from few examples (IEEE TPAMI, 2006) and on Bayesian hierarchical models for natural scene categorization (CVPR, 2005) established foundational frameworks in low-data visual learning that influenced subsequent few-shot and transfer learning literature.
Memoir: The Worlds I See (2023) — Li’s science memoir, published by Flatiron Books, interweaves her personal immigration story with the rise of modern AI. It received wide critical attention and brought her perspective on human-centered AI to a broad general audience.

Awards & Recognition

Queen Elizabeth Prize for Engineering (2025) — Awarded jointly with Yoshua Bengio, Bill Dally, Geoffrey Hinton, John Hopfield, Jensen Huang, and Yann LeCun for foundational contributions to deep learning.
Time Person of the Year — “Architects of AI” (2025) — Named among a small group of figures recognized as defining forces in AI.
Honorary Doctorate, Yale University (2025) — Doctor of Engineering and Technology, with a citation recognizing her as “originator of human-centered AI.”
VinFuture Prize — Grand Prize (2024) — International science prize recognizing transformational contributions to deep learning.
Woodrow Wilson Award, Princeton University (2024) — Princeton’s highest honor for an alumnus in public service.
Elected Member, American Academy of Arts and Sciences (2021)
Elected Member, National Academy of Medicine (2020)
Elected Member, National Academy of Engineering (2020)
ACM Fellow (2018) — Cited for contributions in building large knowledge bases for machine learning and visual understanding.
Intel Lifetime Achievements Innovation Award (2023)
IEEE PAMI Thomas S. Huang Memorial Prize (2022)
IEEE PAMI Longuet-Higgins Prize (2019) — For work of enduring significance in computer vision.
Time 100 AI Most Influential People (2023)
IAPR J.K. Aggarwal Prize (2016)
IEEE PAMI Mark Everingham Prize (2016)
Foreign Policy Leading Global Thinkers (2015)
Alfred P. Sloan Research Fellowship (2011)
NSF CAREER Award (2009)
Microsoft Research New Faculty Fellowship (2006)
Paul & Daisy Soros Fellowship for New Americans (1999)

Academic & Professional Network

Pietro Perona — PhD advisor at Caltech; their collaboration on visual recognition and one-shot learning established the foundational methods Li extended throughout her career.
Andrej Karpathy — Doctoral student; his dissertation Connecting Images and Natural Language (Stanford, 2016) produced key advances in image captioning and visual question answering; went on to found Tesla’s Autopilot AI group and is now at Anthropic.
Timnit Gebru — Doctoral student; became a prominent researcher on algorithmic bias and AI ethics, and a leading voice on the societal dimensions of AI systems.
Olga Russakovsky — Doctoral student; co-organized ILSVRC, now assistant professor at Princeton, and co-founded AI4ALL alongside Li.
John Etchemendy — Co-founder of Stanford HAI; former Stanford provost whose partnership gave the institute institutional standing and cross-disciplinary reach.
Jensen Huang — NVIDIA CEO and co-founder of AI4ALL; his company’s GPUs provided the computational substrate that made ImageNet-scale training feasible; the two have collaborated on AI education initiatives.
Christof Koch — Secondary PhD advisor at Caltech; the cognitive neuroscience perspective Koch brought to Li’s training shaped her lasting interest in bridging machine and human visual systems.
Silvio Savarese — Husband; Stanford professor of computer science whose own research in 3D scene understanding overlaps with Li’s current spatial intelligence agenda.

Personal Style

Li’s intellectual identity is organized around a conviction that AI research divorced from human context produces brittle and potentially harmful systems — a view she translates into institutional structures (HAI, AI4ALL) as much as into technical research agendas. Her public communication is notable for its directness about both the transformative promise of AI and its near-term social risks, a balance she articulated in her 2023 Guardian interview by declining the role of either uncritical advocate or catastrophist. In lectures and writing, she frequently frames scientific questions through biographical narrative, connecting the constraints of her immigrant upbringing — financial precarity, language learning, reliance on exceptional individual mentors — to her commitment to broadening participation in AI. Her TED talks (2015, 2024) and memoir demonstrate a facility for explaining technical ideas to non-specialist audiences without sacrificing scientific substance.