Web native in the '90s.
AI native in the 2020s.

I started in two places at once: Mrs. Livarchik's second-grade Spanish classroom, and Angelfire and GeoCities at home, where view-source was how you learned anything. By sophomore year I'd dropped the host page and was hand-coding a Dave Matthews Band fan site in raw HTML. The Post-Tribune ran a feature on it (Nerdy no more), and I told the paper I hoped the self-taught skills would help while I was "in college and probably very poor." They did. I was.

Music ran underneath all of it: first song written and performed for the class at seven or eight, saxophone in concert and jazz bands through middle and high school, guitar as a teenager, then at sixteen a home-recorded album called Songs Not About John and the open mic nights that came with it. Then the music went quiet for a while, and came back in 2025; I am a strange loop, the song running in the colophon, is one of those.

Both degrees were at Purdue: a B.A. in Spanish, then an M.A. in Linguistics and Second-Language Acquisition. Applied linguistics is the practical side: how people learn, exchange, and build with language. I thought I was studying something academic. It turns out to be the substrate of every product I build now: prompts, specs, onboarding flows, user research as discourse analysis. I thought I was leaving the web behind. I wasn't.

I worked in higher ed for five and a half years before I'd heard the phrase product manager. At Lake Forest College's In the Loop program, I ran residential and academic operations for students living in downtown Chicago for a semester. Human tetris, mostly. On the side, out of habit, I built CASCHA's website and sat on the GLACUHO board, where I learned to cherish a well-run room. Then I met Matt and Dan at Roompact, and the introduction turned higher ed into a software career. Apparently I'd been doing the job for years without the title.

Then came construction payments, of all things. Three and a half years at Levelset, on the team through the $500M acquisition by Procore: lien rights, notarization, e-signatures, regulated document-heavy work that hammers spec discipline in by sheer volume. The lesson that took: boring regulated problems are usually the most valuable ones to solve well. Now I'm at ProNovos as the team's first AI-native product manager, working out what an AI-era product development life cycle looks like in actual practice. Hosette runs alongside it. I'm building on the web again, on purpose.

The braid is language, the web, and music, entwined since childhood, leaving each other for a while, and pulling back together now. Same muscle, new surfaces. Software that takes the domain expert at their word.