Andrew Ng

wiki
, , , ,
1

British-American AI researcher and educator; co-founder of Google Brain and Coursera, Chief Scientist of Baidu, and the most influential figure in AI mass education, having taught over eight million students worldwide.

Basic Information / Profile

Field Details
Full Name Andrew Yan-Tak Ng (吳恩達)
Born April 18, 1976, London, England
Nationality British-American
Current Roles Founder and CEO, DeepLearning.AI; Managing General Partner, AI Fund; Founder, Landing AI; Adjunct Professor, Stanford University; Board Director, Amazon
Research Fields Machine learning, deep learning, reinforcement learning, computer vision, natural language processing, robotics
PhD Advisor Michael I. Jordan
PhD Thesis Shaping and Policy Search in Reinforcement Learning (UC Berkeley, 2002)
Personal Website andrewng.org
X / Twitter @AndrewYNg
GitHub github.com/andrewyng
Google Scholar scholar.google.com/citations?user=mG4imMEAAAAJ
Coursera coursera.org/instructor/andrewng
Newsletter The Batch, via deeplearning.ai

Overview

Andrew Ng is a British-American computer scientist whose career spans foundational AI research, large-scale industrial deployment, and what is arguably the most consequential AI education effort in history. As a Stanford professor he directed the Stanford AI Lab and built CS229 into the university’s most-enrolled course; as a researcher he co-authored the latent Dirichlet allocation paper, led the Stanford Autonomous Helicopter project, and played a central role in laying the technical foundations for Robot Operating System (ROS). In 2011 he co-founded Google Brain with Jeff Dean, and the project’s “cat neuron” experiment — a neural network that learned to recognise cats from unsupervised YouTube video — became a landmark demonstration of large-scale deep learning. As Chief Scientist of Baidu from 2014 to 2017 he built one of China’s most advanced AI research and product divisions. He has since founded or co-founded Coursera (over 148 million learners as of 2024), DeepLearning.AI (more than eight million students), Landing AI, and the AI Fund, a startup studio that has deployed capital across dozens of AI companies. He is among the most outspoken voices arguing that AI existential risk narratives are overstated and that access to AI education and open-source technology is the field’s most pressing obligation.

Early Life & Education

Childhood: London, Hong Kong, Singapore

Ng was born on April 18, 1976, in London. His father, Ronald Paul Ng, was a hematologist and lecturer at UCL Medical School; his mother, Tisa Ho, was an arts administrator at the London Film Festival. Both parents were immigrants from Hong Kong. The family returned to Hong Kong when Ng was young, and in 1984 moved to Singapore. He attended Raffles Institution, one of Singapore’s premier academic secondary schools, before leaving for university in the United States.

Carnegie Mellon University — BSc (1993–1997)

Ng completed an undergraduate degree at Carnegie Mellon University with a triple major in computer science, statistics, and economics, graduating in 1997. Between 1996 and 1998 he conducted research at AT&T Bell Labs on reinforcement learning, model selection, and feature selection — an early sustained engagement with the problems that would define his doctoral work.

Massachusetts Institute of Technology — MS (1997–1998)

At MIT he earned a Master of Science in Electrical Engineering and Computer Science. During this period he built the first publicly accessible, automatically indexed web search engine for machine learning research papers — a precursor to CiteSeerX — demonstrating an early interest in infrastructure that would recur throughout his career.

UC Berkeley — PhD (1998–2002)

Ng completed his doctorate at UC Berkeley under Michael I. Jordan, one of the foundational figures in probabilistic machine learning and statistics. His thesis, Shaping and Policy Search in Reinforcement Learning, examined curriculum-like methods for improving convergence in RL agents and reward shaping techniques. During his PhD he co-authored with David M. Blei and Michael I. Jordan the paper introducing latent Dirichlet allocation (LDA) — a probabilistic topic model that became one of the most-cited papers in machine learning and natural language processing.

Career

Stanford University — Assistant and Associate Professor, SAIL Director (2002–present)

Ng joined Stanford as an assistant professor in 2002 and became associate professor in 2009. He served as Director of the Stanford Artificial Intelligence Laboratory (SAIL). His course CS229: Machine Learning became Stanford’s most-enrolled class, with over 1,000 students registering in some years, and the openly published lecture notes and problem sets became a primary reference for self-taught machine learning practitioners worldwide. His Stanford research group produced several influential lines of work:

The Stanford Autonomous Helicopter project (2004–2008) developed some of the most capable autonomous aerobatic flight demonstrations in the world, using apprenticeship learning to teach helicopters manoeuvres by observing an expert pilot — work that influenced the early formulation of imitation learning in robotics.

The STAIR (Stanford Artificial Intelligence Robot) project resulted in the Robot Operating System (ROS), the open-source robotics middleware that has become the dominant infrastructure for academic and commercial robotics research globally. The project attracted support from Scott Hassan, who subsequently founded Willow Garage to continue the work.

In 2008, Ng’s group was among the first to advocate systematically using GPUs for deep learning training — a technically controversial position at the time that anticipated what became the field’s standard compute paradigm.

He also co-authored formative work on sparse coding, self-taught learning, and multi-task learning, and was the doctoral advisor of several researchers who became major figures in AI: Ian Goodfellow (inventor of GANs), Pieter Abbeel (robotics and RL, UC Berkeley), Quoc V. Le (Google Brain, LLM research), and David Stavens (autonomous vehicles). Ng remains at Stanford as an adjunct professor.

Google Brain — Co-founder (2011–2012)

In 2011, working at Google X with Jeff Dean, Greg Corrado, and Rajat Monga, Ng co-founded what became Google Brain. The project built a neural network trained on 16,000 CPU cores using unsupervised learning on ten million YouTube video frames; without any labelled data or instruction about the concept of a cat, the network developed a neuron that fired strongly for cat faces — a result covered widely in the press as a signal of deep learning’s power. The project’s speech recognition technology was subsequently integrated into Android, and Google Brain grew into one of the world’s most productive AI research labs. Ng left in 2012 to focus on Stanford and Coursera.

Coursera — Co-founder (2012–2014)

In October 2011, Ng and Stanford colleagues launched the machine learning course CS229a online through a custom platform; over 100,000 students registered for the first edition. In 2012, Ng and Daphne Koller co-founded Coursera on the basis of that experiment, with Ng serving as co-CEO. Coursera went on to become one of the world’s leading MOOC platforms; as of 2024 it reported over 148 million registered learners. Ng’s own courses — Machine Learning, Neural Networks and Deep Learning, and AI for Everyone — have consistently ranked among its most popular offerings. He stepped back from Coursera’s day-to-day operations in 2014 upon joining Baidu.

Baidu — Chief Scientist (2014–2017)

In May 2014, Ng joined Baidu as Chief Scientist, relocating to Silicon Valley to lead Baidu’s AI research division — the first major American hire of that scope by a Chinese technology company. He established multiple research teams, including projects on facial recognition, speech recognition, and medical AI (the Melody chatbot). He developed DuerOS, Baidu’s conversational AI platform, and helped position Baidu as the leading Chinese company in AI discourse at a time when the field was being redefined. He resigned in March 2017, citing the desire to pursue broader societal impact beyond a single company’s strategic constraints.

DeepLearning.AI — Founder (2017–present)

In August 2017, Ng launched DeepLearning.AI, a specialisation delivered through Coursera comprising five courses covering neural networks, hyperparameter tuning, structuring ML projects, convolutional networks, and sequence models. The Deep Learning Specialisation became one of the most-taken online courses in any technical discipline, surpassing eight million students. Subsequent DeepLearning.AI curricula have covered MLOps, NLP, TensorFlow, and AI for Good. Ng also publishes The Batch, a weekly newsletter covering AI research and industry, and has distributed Machine Learning Yearning and the AI Transformation Playbook as free practitioner guides.

Landing AI — Founder (2017–present)

Landing AI, founded in 2017, focuses on helping industrial enterprises adopt AI — particularly computer vision and MLOps for manufacturing quality inspection and anomaly detection. The company raised a $57 million Series A led by McRock Capital in November 2021. Its LandingLens platform enables non-ML engineers to build and deploy vision AI applications, reflecting Ng’s long-standing argument that AI adoption requires tools accessible to domain experts rather than machine learning researchers.

AI Fund — Founder and Managing General Partner (2018–present)

In January 2018, Ng launched the AI Fund, initially with $175 million, as an AI-focused venture studio and fund that builds startups from scratch rather than investing in existing companies. The Fund has seeded companies across healthcare, education, logistics, and financial services. In October 2024 it made its first investment in India, backing AI healthcare platform Jivi. Ng has described the AI Fund model as a deliberate attempt to deploy AI research capabilities into verticals where there are few deep ML practitioners.

Amazon Board of Directors (2024–present)

In April 2024, Amazon announced Ng’s appointment to its board of directors, adding an AI research perspective to the company’s oversight at a moment of significant AI investment across AWS and Amazon’s consumer products.

Key Contributions

  • Latent Dirichlet Allocation (LDA) (JMLR, 2003, with Blei and Jordan) — Co-authored one of the two papers that independently introduced LDA, a generative probabilistic topic model that remains one of the most cited works in machine learning and text analysis.
  • Stanford Autonomous Helicopter — Developed apprenticeship learning algorithms enabling helicopters to perform aerobatic manoeuvres through imitation of expert pilots; among the most capable autonomous flight demonstrations of its era.
  • Robot Operating System (ROS) — Led the STAIR project whose technical infrastructure led directly to ROS, now the dominant open-source middleware for robotics research and commercial development globally.
  • GPU advocacy for deep learning (2008) — Among the first academics to systematically push for GPU-based training, a position later validated as the field’s universal compute foundation.
  • Google Brain co-founding and “cat neuron” experiment (2011–2012) — Co-founded Google Brain and led the large-scale unsupervised learning project demonstrating that neural networks could learn high-level visual concepts from unlabelled video; a widely cited signal moment for deep learning’s industrial potential.
  • CS229: Machine Learning — Designed and taught Stanford’s most-enrolled course; the published notes, slides, and problem sets have served as the primary self-study reference for ML practitioners globally for over two decades.
  • Coursera (co-founded 2012, with Daphne Koller) — Co-created the MOOC platform now serving over 148 million learners and offering credentials from over 300 universities and companies.
  • Deep Learning Specialisation / DeepLearning.AI — Designed a curriculum that has enrolled over eight million students, making it the world’s most widely accessed deep learning education programme.
  • AI for Everyone — A non-technical AI literacy course aimed at business leaders and the general public, reflecting Ng’s conviction that AI transformation is as much an organisational as a technical problem.
  • “AI is the new electricity” — A conceptual frame Ng popularised from approximately 2016 onward, arguing that AI’s infrastructural role in the economy parallels electricity’s transformation of every industry a century ago; widely adopted in corporate and policy discussions.
  • Open-source AI advocacy — A consistent public voice against regulatory proposals that would burden open-source model development, opposing California’s SB 1047 in 2024 and articulating the view that licensing requirements for foundation models would concentrate AI capability in large incumbents.
  • Machine Learning Yearning and AI Transformation Playbook — Free practitioner guides distributed directly, each translated into multiple languages and used as onboarding material at AI teams worldwide.

Awards & Recognition

  • Alfred P. Sloan Research Fellowship (2007)
  • MIT Technology Review TR35 Innovators Under 35 (2008)
  • IJCAI Computers and Thought Award (2009) — The highest AI award for researchers under 35.
  • Time 100 Most Influential People (2013, jointly with Daphne Koller for Coursera)
  • Fortune 40 Under 40 (2013)
  • CNN 10: Thinkers (2013)
  • Fast Company Most Creative People in Business (2014)
  • World Economic Forum Young Global Leaders (2015)
  • Time 100 AI Most Influential People (2023)
  • Honorary Fellowship, Royal Statistical Society (2024)

Key Relationships

  • Michael I. Jordan — PhD advisor at UC Berkeley; one of the defining figures in statistical machine learning; the LDA paper co-authored with Jordan remains Ng’s most-cited academic work.
  • Jeff Dean — Google Brain co-founder; led the engineering infrastructure for the “cat neuron” project and the subsequent scale-up of Google’s deep learning research.
  • Daphne Koller — Stanford colleague and Coursera co-founder; the two built the MOOC platform together and were jointly named to Time’s 100 Most Influential People in 2013.
  • Ian Goodfellow — PhD student at Stanford who later invented GANs; in a Mila interview Goodfellow described Ng’s Stanford robotics project and courses as the catalyst for his commitment to AI research.
  • Pieter Abbeel — PhD student who became a leading robotics and reinforcement learning researcher at UC Berkeley and co-founder of Covariant.
  • Quoc V. Le — PhD student who went to Google Brain and led development of seq2seq models, word2vec, and contributions to large language models.
  • Geoffrey Hinton — Intellectual predecessor whose Toronto reading groups influenced many of Ng’s graduate students; Ng’s advocacy for deep learning during Baidu’s peak years ran in parallel with Hinton’s own Toronto-to-Google transition.
  • Baidu leadership (Robin Li) — Under Robin Li, Ng had the resources and mandate to build one of China’s first major industrial AI research programmes, an unusual level of organisational backing for a foreign scientist.

Personal Style

Ng is an unusually explicit educator: where many researchers prefer to communicate through papers, he has consistently chosen pedagogical structures — courses, specialisations, newsletters, free guides — as his primary output format, treating each as a product to be iterated rather than a one-time deliverable. His public position on AI risk is systematic and consistent: he is among the most prominent critics of existential risk framings, arguing that concerns about “killer robots” distract from the concrete labour-market challenges that AI creates in the near term, and that regulatory proposals targeting open-source AI models would harm the researchers and small companies most likely to produce beneficial applications. He has described AI as “the new electricity” and access to AI education as a fundamental equity issue, and his career choices — from free course distribution to free guide publication to a newsletter reaching hundreds of thousands — reflect a genuine commitment to the widening-access position rather than simply a branding strategy.

References