Old-School Chops, New-School AI

I’m Dr. Richard Frnka, and this is the start of a series documenting something I’ve wanted to build for a long time: a real, production-quality mental math game, built almost entirely with AI as my design partner and coding copilot. Before getting into any of that, I want to explain why I’m qualified to try this and, more importantly, why I’m excited about it.

Where this started

I started programming in middle school, in my first computer science class. That turned into a B.S. in computer science and math, with a physics minor, from Eckerd College. From there I went to Louisiana State University for a Ph.D. in mathematics, specializing in analytic number theory.

Grad school wasn’t just research. I taught classes, and I also worked for LSU’s IT department the whole time — so alongside the theory, I was writing and maintaining real software. I also worked with the math department’s capstone course, where undergrad math seniors learned to build small-scale applications that interpreted real data from science projects sourced from other departments across the university. That course is probably the closest thing to a preview of this project I’ve ever done: take people with strong fundamentals and no formal software training, and get them building real, useful tools.

From academia to industry

After grad school, I ran a startup in the Rio Grande Valley, building iOS apps, Android apps, and small websites from scratch. Two years in, I moved to Austin and joined Indeed, where I spent eight years — first as a software developer, then as a product manager.

At Indeed I was something of a hackathon regular, and our last hackathon stands out in particular: it pushed everyone to use AI as aggressively as possible, to design and build as fast as we could actually think. That event changed how I think about the ceiling on solo output. As fast as you can dream something up, you can design and build it with AI — if you know how to drive it.

Old-school chops, new-school AI

That’s really the thesis behind this whole series. I’ve got the old-school chops: two decades of math, a Ph.D.’s worth of rigor, real production software experience across startup and big-tech environments, both writing code and deciding what to build. And I’ve got the new-school AI education from pushing these tools as hard as they’ll go under hackathon pressure.

I want to put both of those to use on something I actually care about.

The game

I’ve been thinking about this particular game for a long time. I had an original version working years ago, built in Swift 2 for iOS. This time, I’m building it for iOS and Android simultaneously, using React Native with Expo and TypeScript, Cursor as my IDE, and AI as both my designer and my coding copilot throughout.

The goal: ship it by December 31, 2026.

Next up: the design process — including everywhere it went wrong before it went right.