What You Get and What You Don't
Khan Academy Linear Algebra is a free collection of video lessons and practice exercises covering vectors, matrices, transformations, eigenvalues, and related topics. It runs inside the Khan Academy website and app, so there's nothing to install or subscribe to. The curriculum follows roughly the same order as a college course, starting with vector basics and moving toward abstract vector spaces. It was designed for high school and early college students, not for professionals who need quick reference material. The videos are short, usually between three and eight minutes each. They're narrated by Sal and other contributors. The pacing is deliberate, which means you'll sit through a lot of setup before getting to the actual computation. That's not a flaw, exactly, but it does mean this resource is better suited for learning from scratch than for reviewing something you already understand. If you just need a refresher on, say, how row reduction works, the videos will feel slow. If you're encountering these concepts for the first time, the step-by-step approach actually helps. I ran into a specific issue while using Khan Academy Linear Algebra to prepare for an interview. The practice problems on matrix determinants use randomized values, but occasionally the system generates a problem where a row is entirely zeros and the matrix is labeled as non-square in a way that doesn't match the problem's intended solution path. The feedback doesn't flag this. My workaround was to verify the matrix dimensions yourself before submitting an answer, and if the problem seems inconsistent, to skip it and move on rather than waste ten minutes debugging the system's logic. This happened to me roughly twice per topic set, so it's not a dealbreaker, but it's worth knowing about.
The exercise interface gives immediate feedback. Correct answers get a green checkmark. Wrong answers show you the correct path after you've made a second attempt, which is useful for building procedural memory but less useful if you're trying to develop intuition about why a method works. The feedback tells you what to do, not why. For people coming from a different background, the notation can be inconsistent. Some videos use column vector notation, others switch to row vector notation without warning. In applied fields like machine learning, you'll encounter row vectors far more often than column vectors in practice, and Khan Academy's default orientation is the mathematical convention. This isn't wrong, but if you're trying to connect the coursework to real tools like NumPy or PyTorch, you'll need to translate mentally. I stopped trying to force that translation during the early lessons and just did it separately afterward. There are also some gaps. The course doesn't cover numerical linear algebra in any meaningful way. If you want to understand why partial pivoting matters in Gaussian elimination or how condition numbers affect computation, you won't find that here. It stops at the theoretical level and treats matrices as objects to manipulate symbolically, not as computational structures with practical constraints. For a complete course, you'd need to supplement with something like Trefethen and Bau or a practical library walkthrough.
The eigenvalues and eigenvectors section is one of the stronger parts. It walks through the characteristic equation, diagonalization, and geometric interpretations with enough worked examples that most students should come away with a solid foundation. The practice problems here are more reliable than in earlier sections, and the difficulty progression feels natural. One thing beginners miss: the course assumes comfort with basic algebra, including manipulating equations and understanding what equality means in different contexts. If you're shaky on solving systems of equations or factoring quadratics, the linear algebra content will feel like it moves too fast, even though the math itself isn't more difficult. The prerequisite knowledge gap is the real bottleneck, not the linear algebra topics themselves. The access is free and the mobile experience works adequately, though the exercise interface is smaller and less responsive than the desktop version. I've seen people struggle with touch input on the matrix entry fields, which can make exercises slower than they need to be. Using a desktop or laptop for the practice problems saves time, especially on the longer topic sets.
Get the Full Details
If you're looking for a structured path through the fundamentals at no cost, Khan Academy Linear Algebra is one of the better options available. It won't replace a full course textbook or prepare you for graduate-level work, but for undergraduate introduction or self-study at the beginning level, it does what it promises without requiring anything beyond an account and a browser.
