Self-taught builder shipping full-stack apps end-to-end — from IndexedDB data layers and serverless functions to AI-assisted product features. Here's a sample of what I've made.
A day-trading options journal and analytics dashboard — my own spin on TradeZella, rebuilt to mirror the workflow I actually trade. Logs every trade, computes options-aware P&L and R-multiples, and generates AI trade reviews. Local-first (IndexedDB) with an optional read-only broker feed.
A daily-use planning tool for people serious about self-improvement — digitally track your goals and plan your days in one place. Layered long-term goals, habit streaks, a lifetime "Freedom Number" milestone tracker, and an AI weekly review, all in one dependency-free HTML file with the AI running server-side on Netlify.
A full-stack sports-betting decision engine. Ingests PropFinder data across MLB/NBA/WNBA, scores plays through an EV model with staking rules, ranks them Elite→Watch, and logs every bet with ROI analytics. Live Kalshi API integration and a Claude co-pilot.
A Streamlit analytics tool for Kalshi's speech "mention" markets — betting on how many times a speaker says a given word. Tracks markets and picks against an edge model (historical-lock / contextual-override / structural-fade), separates event formats, and reports honest ROI with Plotly.
A data-analysis case study: I modeled my own (anonymized) betting log in SQL — star schema, window functions — and built a Power BI dashboard that exposed a 75% win rate hiding a −3.5% ROI, traced the loss to a single pattern, and showed that cutting it flips the book positive (+2.7%). Find the edge, cut the leak.
Every project here I designed, built, and shipped end-to-end — not scaffolded from a template and abandoned. Real architecture decisions (local-first storage, keeping API keys server-side, options-aware financial math), real debugging, and real deployment. I move fast with modern tooling, but I own every decision that ships.