Linear Algebra (18.06)

The lectures explain the intuition, not just the mechanics: why the four subspaces matter, what eigenvalues actually mean, how SVD ties it together. That understanding is what separates people who can debug an ML model from people who only run one. It rewards working through the problems rather than just watching.

It is pure mathematics, so there is no coding and no direct ML application here, and no certificate. The recordings are old and the production shows it, but linear algebra remains relevant and the teaching still beats most modern alternatives. Pair it with a hands-on ML course to connect the theory to code.

MIT 18.06 with Gilbert Strang is the most-watched university math course in history — Strang's lectures have been viewed by an estimated 4 million learners worldwide, and his textbook defines how linear algebra is taught at the undergraduate level globally. The course covers vector spaces, matrix operations, systems of linear equations, determinants, eigenvalues and eigenvectors, the four fundamental subspaces, and the singular value decomposition (SVD).