Andrew Ng

Who they are

Andrew Ng is Managing General Partner at AI Fund — co-founded Google Brain in 2011, Coursera in 2012, and DeepLearning.AI in 2017, and recently shipped OpenWorker, an open-source AI agent built with Rohit Prasad.

Person

Andrew started AI Fund in 2017 when he bootstrapped DeepLearning.AI from scratch — no outside capital, self-funded from day one. He came up through Carnegie Mellon (BS in Math, Computer Science, Statistics, and Economics), then MIT (MS in Electrical Engineering and Computer Science), then UC Berkeley (PhD in Computer Science, 2003) — a research pedigree that shaped everything after. At Stanford he ran the AI Lab and led the Autonomous Helicopter and STAIR projects, the latter producing ROS (the Robot Operating System). He co-founded Google Brain in 2011, pioneering large-scale neural networks that predated TensorFlow, then left for Baidu as Chief Scientist and VP before going independent in 2017. The companies he's started tell the story cleanly: Google Brain (deep learning at scale inside a major tech company), Coursera (online education platform, co-founded with Daphne Koller in 2012, now over 100 million learners), DeepLearning.AI (AI education and The Batch newsletter, founded 2017), Landing AI (industrial AI and Document AI, 2017), AI Fund ($175 million venture studio co-founding AI startups, 2018), and most recently OpenWorker (open-source AI agent that delivers finished work — documents, Slack messages, calendar updates — while preserving data privacy, co-developed with Rohit Prasad). The through-line is consistent: make AI accessible and deployable, whether through education, research infrastructure, or applied tooling. He publishes The Batch weekly at DeepLearning.AI covering AI news, LLMs, and industry implications, writes at andrewng.org, posts actively on LinkedIn about agentic AI, open-source tools, and AI policy, and has appeared at Ai4 2026 alongside Geoffrey Hinton and Fei-Fei Li.

Company

DeepLearning.AI, which Ng founded in 2017 as a bootstrapped, self-funded company, reported $52.8 million in annual recurring revenue and a disclosed valuation of $158.4 million as of 2025 — no outside venture capital has been raised. The most recent product move is a 2026 public GitHub repository releasing developer tools, course artifacts, and a master course catalog. Earlier, in January 2024, the platform launched a free short course on automating LLM evaluations in partnership with CircleCI. Course partnerships with AWS, Google, OpenAI, and Red Hat anchor the curriculum, and the platform collaborates with Sharon Zhou on offerings like Transformers in Practice.

Market

DeepLearning.AI competes in AI education and technical training against ai.google, D2L.ai, Pluralsight, Codecademy, and NxtWave, while sitting in the broader orbit of major AI platform providers like OpenAI and Anthropic. The AI education market is shaped by rapid model releases that constantly refresh what practitioners need to learn, keeping course content on a short shelf life. Globally, regulatory divergence — the EU's stricter AI frameworks versus the US's market-driven approach — and intensifying US-China geopolitical competition over AI technology increasingly influence where companies like DeepLearning.AI can partner, distribute, and recruit.

Network

Ng's closest named collaborator on recent work is Rohit Prasad, with whom he co-developed OpenWorker. His longest-standing academic and industry peer relationships include Daphne Koller (Coursera co-founder), Fei-Fei Li (co-panelist at Stanford HAI and Ai4 2026), and Geoffrey Hinton (co-panelist at Ai4 2026). On the enterprise side, he has active relationships with Morgan Stanley (AI Challenge Lab, with Daniel Simkowitz, Andy Saperstein, Sharon Yeshaya, Matt Berke, Cameron Rice, and Adam Kudelka), Bob Bradway of Amgen (AI and pharma conversation, 2026), and Bain & Company through a strategic AI transformation partnership.

Rohit Prasad· Co-developer, OpenWorker
Daphne Koller· Co-founder, Coursera
Fei-Fei Li· Co-panelist, Stanford HAI and Ai4 2026
Geoffrey Hinton· Co-panelist, Ai4 2026
Sharon Zhou· Course collaborator, DeepLearning.AI
Bob Bradway· CEO, Amgen — interviewed on AI and pharma (2026)

How they likely show up

Founder role type across six separate companies (Google Brain, Coursera, Landing AI, DeepLearning.AI, AI Fund, OpenWorker) → he defaults to starting things rather than joining them; expect high initiative and low patience for bureaucratic drag.
DeepLearning.AI has been bootstrapped since 2017 with no outside funding → he runs lean and is comfortable with long time horizons without external pressure milestones.
Active LinkedIn poster on agentic AI, open-source tools, LLMs, and AI policy → comfortable being a public voice; responds well to engagement with specific ideas he's published, not just general flattery.
Publishes Machine Learning Yearning and AI Transformation Playbook for free, runs The Batch weekly → education and knowledge-sharing are intrinsic, not promotional; he treats access to information as a first principle.
Co-developed OpenWorker with Rohit Prasad as an open-source project alongside running AI Fund → high agency, builds in public, likely bored by closed-loop internal strategy that doesn't ship anything.
Research background (Stanford AI Lab, Google Brain, Baidu Chief Scientist) combined with operator track (Coursera Co-CEO, AI Fund GP) → can move between technical depth and business abstraction without needing a translator.

Conversation tips

→ Reference a specific The Batch letter or a recent LinkedIn post — he'll know immediately whether you've actually read it or are citing it by title alone.
→ Ask about OpenWorker specifically: it's a recent open-source bet on privacy-preserving agentic work, and it signals where he thinks the next inflection is.
→ If you're in education, healthcare, or pharma, name that context early — his public calendar shows he's actively engaging pharma leaders (Bob Bradway) and healthcare AI (Stanford HAI), so domain specificity will land.
→ Don't frame questions at the level of 'AI in general' — he operates at the level of specific deployment patterns (agentic coding, LLM evaluation, industrial AI); match that specificity.
→ The Bain partnership and Morgan Stanley AI Challenge Lab are recent enterprise moves worth acknowledging if you're in professional services or financial services — they signal where he's taking AI Fund portfolio thinking.

Toolbox

Openers

Open on OpenWorker — he recently co-built an open-source AI agent with Rohit Prasad that delivers finished work (documents, Slack messages, calendar updates) while keeping data on-device; it's a pointed bet that agentic AI should be open and privacy-preserving, worth unpacking why he went open-source rather than commercial.
Reference the Ai4 2026 keynote panel with Geoffrey Hinton and Fei-Fei Li — three of the most consequential figures in modern AI on one stage is an unusual signal about where the field's founders think the public conversation needs to go right now.
Bring up the DeepLearning.AI bootstrapped path: $52.8 million ARR and a $158.4 million valuation with zero outside funding since 2017 is a deliberate structural choice from someone who could raise at any valuation — the decision to stay independent is worth exploring directly.

Discovery questions

OpenWorker ships finished work as an open-source agent — what made you choose open-source over building it inside AI Fund as a commercial product?
DeepLearning.AI has stayed bootstrapped while course partnerships with OpenAI, AWS, and Google keep the curriculum current — how do you manage curriculum half-life as models release faster than courses can be updated?
You've co-founded companies across research infrastructure, education, industrial AI, and now agentic tooling — where does AI Fund focus its company-building energy in 2026, and what problem shape are you most actively looking to start next?

Avoid

Don't open with broad claims about 'the AI revolution' or reference his background as simply 'the Coursera guy' — he operates at high technical and strategic specificity, and generic framing signals you haven't engaged with what he's actually building now.

Make it yours