Projects
An assortment of projects I have played around with over time - often to try out a new technology, or to fill a need I couldn't find a good solution for on the market, and found it easier to build myself.
MathBop
A math app I built to keep my daughter engaged over the summer so her math stays fresh. I didn't like Duolingo's math courses, so I made something in the same style, but more geared toward simple repetition, in the spirit of Kumon. Ideal for ages 5-7. And it worked - my daughter got hooked. I threw in some gamification by letting you unlock new versions of the sandwich mascot as you complete more levels.
JazzUp
This is the jazz instructional app I always wished existed. To really learn jazz, you either need a great teacher, or you memorize standards and then apply theory to build your own improvisations. But there's no interactive guide to help you develop an improvisational repertoire. JazzUp dynamically adjusts the chords and melody with increasing levels of complexity, so you can pick up new comps while following along.
Floor Planner
While moving, I wanted a simple way to assess how furniture would fit and be laid out. I tried tools like Lucid and GIMP, and explored some free options online, but they were all either clunky or overkill. This lightweight tool lets you start with a floor plan image as your background and overlay various shaped rectangles over it. Dead simple, with accurate calibration for real dimensions.
Validaus Validaus
A side project I built in 2018 that drew strong early interest from friends and colleagues. The core problem it tackled: your regression tests in UAT may all pass, but once you're running in production, how do you actually know your data is correct and matching the business's rules? Validaus let non-engineers define those rules, which we called “Facts,” in a no-code interface, connect their data sources, and get automated quality reports and alerts without touching the codebase. It gave developers, product managers, and analysts a shared source of truth for what correct data really means. The idea was to provide what I called “Business Coverage,” as opposed to “Code Coverage”: you could call a single method, like pricing a claim, with a whole set of real-world inputs and confirm that the outputs matched your expectations.