The Complete Stanford AI & ML Lecture Collection

Andrew Ng's legendary ML course. Linear regression, logistic regression, SVMs, neural networks, unsupervised learning, EM algorithm. The foundation every ML engineer needs before anything else.

CNNs, image classification, object detection, segmentation, generative models, and the full visual recognition pipeline. Taught by Fei-Fei Li and team — the definitive vision DL course.

Word embeddings, RNNs, attention, the transformer architecture, pretraining (BERT, GPT), and modern NLP systems. If you want to understand how LLMs work from the ground up — start here.

Graph neural networks, node embeddings, knowledge graphs, GraphSAGE, GCNs, and applications in social networks, biology, and recommendation systems. Essential for Graph RAG and knowledge graph work.

How models learn to learn. MAML, Reptile, few-shot learning, transfer learning, multi-task architectures, and the techniques behind models that generalize from very little data. Taught by Chelsea Finn.

MDPs, Q-learning, policy gradients, actor-critic methods, deep RL, and exploration strategies. The math and intuition behind training agents that learn from reward — including the foundations of RLHF.

Guest lectures from researchers at OpenAI, DeepMind, Google Brain, and Anthropic on cutting-edge transformer applications — vision, language, robotics, science, and beyond. Closest you get to a frontier research seminar for free.

Broad foundations of AI — search, constraint satisfaction, Bayesian networks, Markov decision processes, logic, and learning. The course that builds the conceptual vocabulary every AI practitioner needs.

Second offering of Stanford's graph ML course with updated content, new guest lectures, and deeper dives into heterogeneous graphs and scalable GNN training. Different instructors, different angle on the same powerful ideas.

Stanford's dedicated LLM course. Scaling laws, emergent capabilities, alignment, RLHF, societal impact, hallucination, and what frontier models actually do under the hood. Most directly relevant to where AI is today.