Notes from someone learning in public.

I study deep learning and probability, and I build small, interactive instruments that make abstract structure easier to inspect. This site is the workshop: a title page, a few deep dives, a running research log, and a set of numbered plates you can actually touch.

The throughline is the same in all of it — to understand a thing by making it move. A prime spiral, a strange attractor, a network guessing a digit you drew. The visuals are not decoration; they are how I think, set down so someone else can think alongside.

What I am working through

• Training — optimization, inference, and how architectures actually behave.
• Probability — structure, uncertainty, and building real mathematical intuition.
• Interfaces — simulations that explain by moving rather than by telling.

Correspondence

The most reliable way to reach me right now is GitHub, where the source for this site and its plates lives.