Projects

Personal project

Situation

An older ASP.NET MVC beer catalog app needed a modern rebuild, and once that rebuild proved out a staff-confirmed loyalty flow for one business, the real question became whether the same pattern could work for any business that wants members working through a list of things to earn a reward, without a rebuild for each one.

Action

Rebuilt the app with a React frontend, an ASP.NET Core backend, PostgreSQL, and Docker; then added real multi-tenancy, per-business configuration (goals, item names, badge tiers, branding), and the AWS infrastructure to run many differently-configured businesses from one codebase.

Result

A live-tested platform running five fully isolated, differently branded demo businesses (a bar, a coffee shop, a brewery, a record store, and a dispensary) from the same code, deployed and verified end-to-end on real AWS infrastructure, with a Redis-backed cache layer for running more than one instance and the first industry-specific bespoke redesign now underway.

Self-initiated

Situation

A resume and LinkedIn profile can describe the combination of PM and development experience, but they can't demonstrate it directly.

Action

Scoped and shipped this site incrementally on Next.js — home, resume, and blog first, then a database-backed comment system with Neon Postgres and Cloudflare Turnstile bot protection — the same way I'd break down and sequence any real project.

Result

A live, evolving site where each new feature is shipped as its own scoped unit of work and documented on the blog so others can build the same thing.

Personal project

Situation

I wanted an edge in my own fantasy football league — not generic rankings, but a system built from this league's actual scoring rules, draft history, and the tendencies of the specific people I draft against.

Action

Built an AI-powered draft and season assistant on top of 16 years of real NFL data (nflverse), scored under this league's exact rules rather than generic PPR assumptions. It computes value-over-replacement tiers from the league's own scoring history, tracks coach and scheme tendencies from real play-by-play data, and cross-checks rankings across ESPN, Yahoo, and CBS to spot which players the room is likely to undervalue. Connected live to the ESPN league API and delivered through a set of Claude Code skills, it turns all of that into a tiered draft board, real-time pick tracking with reach/value alerts, and — new this week — a weekly waiver-wire assistant.

Result

Used it live to draft an entire 12-team keeper league roster pick-by-pick, catching real-time value plays other managers missed. Now expanding into weekly in-season decisions as the 2026 season kicks off.