FEI-FEI LI

Fei-Fei Li was born in Beijing in 1976. Her family immigrated to the United States when she was 16, settling in Parsippany, New Jersey.

Her parents worked in a dry cleaning shop and a restaurant. She got into Princeton on a full scholarship, studied physics, and graduated in 1999.

Then she went to Caltech for a PhD in electrical engineering, finishing in 2005.

Her early academic work was in computer vision — teaching computers to recognize objects in images. The field had a problem: there wasn't enough data.

Researchers were training models on small, carefully curated datasets that looked nothing like the real world. Li's insight was simple and turned out to be enormous: make the dataset the size of the real world.

In 2006, she started building ImageNet. She scraped the internet for images, organized them into 22,000 categories using WordNet's semantic hierarchy, and hired workers through Amazon Mechanical Turk to label each image.

By 2009, ImageNet had 14 million labeled images across 20,000 categories. She launched the ImageNet Large Scale Visual Recognition Challenge (ILSVRC) as an annual benchmark.

For three years, the results were incremental. Then in 2012, a team from the University of Toronto entered a deep neural network called AlexNet.

It beat the second-place system by a 10.8 percentage point margin — which was not a small improvement, it was a seismic one. The field realized the approach was fundamentally different.

Deep learning had arrived. ImageNet was the fuel.

Li joined Stanford's Computer Science faculty in 2009, where she has been a full professor since. She took a leave of absence from 2017 to 2018 to serve as Chief Scientist of AI/ML at Google Cloud.

She returned to Stanford and co-founded the Stanford Institute for Human-Centered AI (HAI) in 2019. In 2023, she published her memoir 'The Worlds I See.' In 2024, she left Stanford on leave to found World Labs, a spatial intelligence AI startup that raised $230 million at a $1 billion valuation.

World Labs (co-founded 2024) is Li's most recent and most commercially ambitious venture. The company focuses on spatial intelligence — teaching AI to understand three-dimensional space, physical environments, and how objects interact in the real world.

This is the dimension of intelligence that current large language models lack. World Labs raised $230 million in its seed round from investors including Andreessen Horowitz, Radical Ventures, and others.

The valuation hit $1 billion before the company had launched a product.

Stanford HAI (co-founded 2019) is the research institute Li built at Stanford to focus on AI that benefits humanity. It produces policy research, funds interdisciplinary AI work, and has become the most credible academic voice on AI governance in the US.

It's not a commercial entity — but it has shaped how governments and companies think about AI safety and ethics more than most commercial efforts.

AI4ALL (co-founded 2017) is a nonprofit Li created to increase diversity in AI education. It runs camps and programs for underrepresented high school students, giving them access to AI training that would otherwise require attending MIT or Stanford.

Since its founding, it has reached over 6,000 students across 26 universities.

Li doesn't invest in the traditional sense — she builds. Her approach is research-first, long-term, and mission-driven.

She identifies fundamental problems in AI capability (not enough labeled data in 2006, not enough spatial understanding in 2024) and builds infrastructure to solve them. She's not trading equities or backing portfolio companies through a fund.

She's placing decade-long bets on what AI needs to become.

The pattern is consistent: find the bottleneck, build the thing that removes it, release it to the world if it's research, build a company if it's commercial. ImageNet was the first move.

World Labs is the latest. The time between was spent ensuring the field developed in a way that includes rather than excludes.

ImageNet. Not because it made her rich — it was open source and free — but because it changed everything.

The 2012 AlexNet result on her dataset triggered a decade-long AI arms race that reshaped technology, created trillions of dollars in market value, and put AI at the center of every consequential industry on earth. She didn't do it for the outcome.

She did it because she thought computer vision needed better training data. She was right in a way that turned out to matter more than anyone predicted.

That's the most interesting kind of right.

Li has been candid about the tension in her time at Google Cloud (2017–2018). She went to help shape how one of the world's most powerful companies uses AI.

She came back with complicated feelings about the difficulty of influencing AI development from inside a corporation at that scale. The specific controversy was over Google's Project Maven — a military AI project — which she reportedly had concerns about.

The internal debate about AI and national security was real and ongoing. She returned to Stanford and refocused on research and policy.

The lesson, if there is one: institutional power is hard to steer from inside when you're not the one driving.

Li's financial philosophy comes from someone who started with nothing. She has spoken about her family arriving in the US with limited means, her parents working multiple jobs, and her scholarship to Princeton being the thing that opened the door.

Her view on wealth is instrumental: money matters because of what it enables — research, access, opportunity for others. She doesn't fetishize it or avoid it.

She has been direct about the fact that AI creates enormous economic value and that the people building it deserve to participate in that value. World Labs is not a nonprofit.

She is building a commercial company. That's the evolution from her earlier career: believing you can build something that matters and that also generates returns is not a contradiction.

Li is married to Silvio Savarese, a computer vision researcher and professor at Stanford. They met through their shared field and have two sons together.

Her parents immigrated from China and worked in service jobs in New Jersey — that origin is one she references often, because it informs her view of what opportunity looks like when it's absent and what it costs to be excluded from the systems that create it. She has spoken about raising her children with awareness of the technology landscape they're growing up inside.

Princeton University, BA in Physics, 1999. Caltech, PhD in Electrical Engineering, 2005.

Her Princeton thesis was in physics but her mind was already moving toward computation and intelligence. At Caltech she found the research direction that would define her career.

The physics training shows — her approach to problems is to find the underlying structure, not just optimize the surface.

As an Amazon Associate, Netfigo earns from qualifying purchases. Book links above may be affiliate links.