0 of 14 done
The math that underpins machine learning and deep learning — shared by the Data Scientist and ML Engineer paths (the AI Engineer path needs far less of it). Part of 00-foundations.
Each area is deep enough to be its own folder of atomic notes rather than one flat file.
Reading order
- linear-algebra/ — [Beginner → Intermediate] — vectors, matrices, dot products, matrix multiplication — why weights and embeddings are what they are
- calculus/ — [Intermediate] — derivatives, gradients, the chain rule — why backpropagation works
- probability-and-statistics/ — [Beginner → Intermediate] — distributions, expectation/variance, Bayes — why loss functions and sampling look the way they do
- optimization — [Intermediate] — gradient descent, learning rate, batches — how training actually runs
Gaps to fill later (per roadmap.sh)
The Data Scientist and ML Engineer roadmaps expect more than the above: more linear algebra (eigenvalues/eigenvectors, SVD, determinants & matrix inverse), inferential statistics (hypothesis testing, confidence intervals, distribution types) beyond the descriptive basics, and discrete math. These are noted for a later pass.
Related
- Numerical Methods — conditioning, why gradient descent zigzags, and the linear algebra done properly
- Information Theory — where cross-entropy loss and KL divergence actually come from
- Foundations
- AI/ML course map
Jump to a subject
Software engineering
Infrastructure & security
Computer science
Hardware & engineering