Where curiosity meets impatience.

Spotify as a Behavioral Dataset

A Python pipeline that treats my Spotify library as panel data: playlist embeddings, genre drift over time, monthly consumption dynamics. It started as curiosity, turned into a crash course on clustering and scripting myself, and now doubles as a diary of how taste shifts under stress.

Financial ML Teaching Notebooks

Reusable Jupyter notebooks that try to make the math feel grounded:

• Visual bias–variance tradeoff
• Regularization as a form of belief‑shrinkage
• How factor models sneak into ML regressions

Built while TA‑ing at Chicago Booth and iterated in real classrooms. The goal is to leave students with code they can actually remix.