Lecture notes, written to be reread
Study
Stanford computer science courses worked through from the public lectures. One page per course, one section per lecture. Equations included, assignment solutions not.
Machine Learning
Supervised learning from least squares to SVMs, learning theory, EM, PCA, and MDPs, with the equations that tie them together.
17 lectures →CS230 · PracticeDeep Learning
Applied deep learning from the forward pass to shipping a project: optimization, CNNs, detection, sequence models, GANs, and project strategy.
15 lectures →CS231n · Computer visionDeep Learning for Computer Vision
From pixels to ResNets, ViTs, detectors and generative models: how modern vision systems are built, trained and understood.
17 lectures →CS224n · LanguageNatural Language Processing with Deep Learning
Study notes on CS224n: word vectors, RNNs, attention, Transformers, pretraining, prompting, LLMs, and the analysis and ethics questions around them.
17 lectures →CS224W · GraphsMachine Learning with Graphs
Notes on CS224W: node embeddings, PageRank, GCN, GraphSAGE, GAT, GNN expressiveness, knowledge graphs, recommenders, and scaling.
16 lectures →CS234 · DecisionsReinforcement Learning
Study notes on learning to act from reward: MDPs, TD and Q-learning, DQN, policy gradients, exploration, offline RL, MCTS, and RLHF.
15 lectures →CS246 · ScaleMining Massive Data Sets
Algorithms that still work when the data does not fit in memory: hashing, sketches, sampling, and linear algebra at cluster scale.
15 lectures →CS336 · SystemsLanguage Modeling from Scratch
Study notes on building a language model end to end: tokenizer, Transformer, training systems, scaling laws, data, inference and alignment.
15 lectures →Personal study notes written in my own words from publicly available lectures and course pages. Not affiliated with or endorsed by Stanford University. No assignment solutions are published here. Course names and lecture order follow the official course pages, linked from each note. If you spot an error, open an issue.