Building a labor-of-love mobile app for tracking vessels with the Washington State Ferry system. First built as a prototype using React, as an intended mobile Progressive Web App, after "feeling the pain" of living on Bainbridge Island for years with frequently-late ferries. (Back in Seattle today.)Now being redeveloped with the following stack for cross-platform mobile and web-apps:▪ Teaching self React Native using the Expo framework for a "three screens" experience for iOS, Android, and the Web, including submission to app stores when ready.▪ Functional programming techniques with TypeScript (JavaScript) for cleaner code and fewer bugs.▪ Real-time vessel positions using REST/JSON calls to a WSF API server, with data ever ten seconds, with data smoothing to look less janky.▪ Beautiful, performant, customizable 3D maps using the Mapbox platform for eye candy.▪ A serverless backend in Google Firebase with serverless functions to ping the API server for locations and persist historical location data to a NoSQL database. (Contemplating a SQL alternative since the data is highly-structured, whether through a Docker container on some micro instance, or a serverless SQL provider.)▪ Mobile-first design using the Tamagui CSS-in-JS framework, roughly like Tailwind, including support for animations and other "nice to have" graphical embellishements. (Helpful for those of us wtihout real design skills.)▪ Anticipated "freemium" model around "free" for infrequent users, "lower cost" for regular walk-on passengers (e.g., real-time schedule alerts about always-late ferries), and "premium" for drivers who already pay through $50 round trip (including using "spaces remaining" API data with a simiple AI to predict "when to get there.") Additional opportunities for sponsorships by ferry-adjacent businesses, to come.▪ Intended as a side-hustle weekend project, not as a full-time position.

Traded own brokerage account using a commercial algorithmic stock-trading IDE (MultiCharts) whle attending school. The stack involved a writing simple scripts at the level of "buy if these triggers are met," developing and backtesting models around historical data from a data-broker, deploying strategies to a rented server, and making supervised (later unsupervised) trades to stock-broker via API. Had good results for several years, got overconfident, and found that the house always wins.

As a hobby developer and law student, conceived of a Microsoft Word plugin for checking legal briefs for the expected citation styles, subject to many rules, like a grammar checker. Leared C# from books (which still existed), learned to build UIs using Microft's UI library, and built a rudimentary prototype as a plugin for Microsoft Word. Then (without fully appreciating the implications), ran into brick walls of parsing ambiguous natural language wile finding errors, roughly like "trying to find and correct malformed French sentences in an English document." Was awared U.S. Patent 7,028,259, "Interactive Legal Citation Checker" in 2006, but not commercialized.