Websites and web apps, built and run by one person.

Look up any player on any team and see their season — game by game, with snap counts and a chart of how their year actually went. Behind it is a pipeline that pulls play-by-play, rosters, and injury reports from public NFL data on a schedule, loads them into Postgres, and records whether each run succeeded. Nothing is scraped when you load the page. Four services across three hosts, and none of them is a computer I own.

A dark, atmospheric six-page site for a tattoo studio — services, pricing, a large photo gallery, and online booking — with custom typography and a visual identity built from scratch.

A clean, mobile-friendly site for a neighborhood vacuum sales-and-repair shop, with an AJAX contact form so customers can reach out without leaving the page.

A self-service booking tool for an appointment business. Customers pick a service, day, and open time slot; the system prevents double-booking and updates availability in real time. A built-in owner dashboard shows every booking, blocks off time, and manages services — no code required.

A team wellness dashboard that tracks player load, daily wellness check-ins, and injury history across a roster. Coaches pick a team, drill into any player, and see training-load and wellness trends charted over time — backed by a seven-table PostgreSQL schema and a typed REST API.

A database-driven dashboard for tracking patient health data, with appointments, vitals trends, medications, lab results, and messaging — a live demonstration of dynamic, data-backed web software.

A meal-planning app on a REST API with a normalized six-table database. Scale any recipe and every ingredient recalculates live; add several recipes and it merges them into one shopping list, converting units and grouping by aisle.

A 12-team mock draft you can actually play, with machine-learning projections behind every pick. Models trained on fifteen seasons of real NFL data give each player a floor, median, and ceiling instead of one flat number — so you can see who has genuine upside, not just who scored well last year. Eleven AI opponents draft off live market data pulled from real drafts happening right now.

A week-by-week touchdown and receiving-yards forecast for one NFL receiver — and an honest account of what happens when you test a model properly instead of trusting it. Cross-validation showed the model barely beats a simple average, so the write-up says exactly that, documents the features that got tested and cut, and keeps judgment calls visibly separate from what the data actually supports.