Start here
Everything on this site is free and self-paced: 12 modules that take you from linear regression to transformers and generative AI. This page tells you what to know before Module 1 and how to get the most out of the course.
- Calculus: multivariable derivatives, the chain rule, and gradient computation
- Linear algebra: vectors, matrices, eigenvalues, rank, and least squares
- Probability & statistics: distributions, expectations, maximum likelihood estimation
- Programming: Python fluency with NumPy; comfort writing and debugging functions
- ML fundamentals: supervised vs. unsupervised learning, overfitting, train/test splits
Rusty on any of these?
Spend a few days with these free refreshers before starting — Module 1 will feel much smoother:
Every module page is a complete, self-contained hub. Work through it in this order:
- 1Watch — the module's lecture videos (most are board lectures — have pen and paper ready).
- 2Read — the linked course-book chapters, interludes, and reference appendices in the listed order.
- 3Code — open the module's Colab notebook and run/modify the lecture code yourself.
The free course book is the canonical reading. Its current route spans all 12 modules and pairs core chapters and interludes with the reference appendices needed for each topic. Slides and primary-source papers remain available when a second treatment is useful.
Suggested pacing: budget 5–7 focused hours per module, roughly one module every 1–2 weeks. The modules build on each other — resist skipping ahead unless the prerequisites note on a module page says you're clear.
The fastest path is Google Colab — free GPU access, nothing to install, and every module's notebook opens directly in it. Prefer working locally? The setup guide covers Python, PyTorch, and Jupyter.