Teaching

I teach matrix theory and matrix computations to graduate students in AI. The organising idea is that numerical analysis is not background material for machine learning — it is the part that decides whether a method works at scale and in floating point.

Matrix Theory and Applications I

M571011S03 Graduate 16 weeks · 32 contact hours 2026–

A graduate course on matrix theory and matrix computations for students in artificial intelligence and related disciplines, built on Golub & Van Loan's Matrix Computations.

Runs the book's own build order rather than the conventional one: the SVD arrives in Week 3 as an analysis tool, and its stable computation waits until Week 14, after QR iteration makes it derivable.
Conditioning is taught in Week 4, before any solver, so every algorithm that follows can be judged against it.
Five programming assignments, each following the same rule — implement it, then verify against a reference implementation.

Course website → — full schedule, readings by section, assignments, and slides.