A curated selection. Not everything, just the work I want to show.
An end-to-end ML system that predicts ATP/WTA tennis matches, finds games the betting market has priced wrong, and runs the whole loop — data, model, value detection, bankroll tracking — unattended. Built solo, front to back.
A cross platform program for music track metadata organization. Special capability - automatically populates track metadata using audio fingerprinting. I ran it as a SaaS before shutting it down; now open-sourcing it.
My data company. Python pipelines aggregate 3.8M+ public records from 16 government sources — via APIs, undocumented endpoints, and browser automation — into one normalized schema served through clean REST APIs.
Automated trading algorithms (Expert Advisors) written in MQL for MetaTrader — built before generative AI was public, back when I was teaching myself to code by shipping things that had to actually work.