Seven interactive primers on the math, optics, and models that make machines see. Drag the sliders. Break the inputs. Watch eigenvectors track a point cloud, disparity collapse with depth, and CLIP rank images by typed text. Built as I learn — corrections welcome.
Each post is a single self-contained HTML file. The visualizers run on raw 2D canvas — no D3, no react, no build step. Source for everything is in the portfolio repo. If a viz doesn't make sense or the math is off, tell me; I'd rather fix it than leave it wrong.