Courses
Structured curricula if you want a path rather than browsing topics. From gentle introductions to grad-level specialty courses.
Introductions
Start here if ML is new to you. Pick one — don't try to take them all at once.
- Andrew Ng — Machine Learning Specialization The canonical "first ML course". Audit free, certificate paid. Solid foundations on classical algorithms; light on deep learning.
- fast.ai — Practical Deep Learning for Coders Top-down, hands-on deep learning. You build state-of-the-art models in week 1. Best complement to a theory-first course like Andrew Ng's.
Stanford classics
The gold-standard university courses, all with public materials.
- CS229 — Machine Learning Stanford's flagship ML course. Lecture notes are widely considered the best concise ML textbook ever assembled.
- CS231n — CNNs for Visual Recognition The gold-standard deep-vision course. Slides and assignments are public; the lecture videos cover the foundations of CNN training.
- CS224n — NLP with Deep Learning Embeddings → attention → transformers. Updated yearly. Lecture videos on YouTube; assignments are public and substantive.
- CS236 — Deep Generative Models Stefano Ermon's course. The structured introduction to VAEs, GANs, flows, diffusion. Public lecture notes are the best single reference for the area.
Reinforcement learning
- CS285 — Deep RL (Berkeley) Sergey Levine's grad-level RL course. The reference for understanding modern RL algorithms — policy gradients, actor-critic, off-policy methods, model-based RL.
- David Silver — Introduction to RL DeepMind's David Silver teaching the foundations. Friendlier than CS285; best paired with Sutton & Barto's book.
Coming soon
Additional courses I want to add — MLOps, recommender systems, multimodal, foundation models.