For most of my career, the work I did lived inside spreadsheets, models, and deal rooms no one outside the firm would ever see. I'd spend weeks underwriting a portfolio and the output was a committee deck that disappeared into a folder forever.
So I started building things with a URL instead.
This is the first entry in a build log I'll keep going as I ship. No thought leadership, no fluff — just what I'm building, the decisions behind it, and what I'd do differently. If that's your thing, subscribe at the bottom and you'll get each dispatch in your inbox.
The through-line
Everything here follows the same loop: I run into a problem I actually have, I build the tool that solves it, and I ship it with a real address on the internet. None of these started as "products." They started as things I needed.
The other thread is how they get built. I work alongside AI — Claude Code, specifically — as a force multiplier, which is how one finance guy ships this much software on nights and weekends. More on that below.
HPI Command Center
The first thing I built that mattered was an internal underwriting system for commercial real estate. The short version: it takes the deal analysis and portfolio monitoring that used to eat the better part of an hour and does it end-to-end in a few minutes.
It underwrites deals, keeps a live picture of the portfolio, and turns a pile of disconnected data into something you can actually make decisions from. I built it in-house, on my own initiative, using the same AI-assisted workflow I use for everything else. It's not a side project — it's a daily driver at work.
Wafergraph
Wafergraph is the semiconductor and AI supply-chain mapped as a graph. The whole value chain — from raw materials and equipment, through EDA, IP, chip design, foundries, memory, packaging, and distribution, all the way to the AI and data-center layer — laid out so you can actually see how it connects.
It tracks 450+ companies across a dozen value-chain segments, with financials, who depends on whom, market share, and where the real chokepoints are. If you've ever wanted to understand why one fab or one materials supplier can hold up the entire AI buildout, this is the map.
Who's Starting
Who's Starting is the most "scratch my own itch" thing on this list. It's a mobile-first college football depth-chart viewer built for game day — open it on your phone, pick the matchup, and see who's actually starting, fast, without fighting a desktop site that wasn't built for the couch.
Under the hood it pulls live depth charts and team identity so it stays current through the season. It's built for this fall.
The site itself (and this log)
The thing you're reading this on is also a build: a live résumé that doubles as a place to show off what I make, with an operator-terminal aesthetic because I think portfolios should be fun. The build log you're reading is the newest piece of it.
I also open-sourced the Claude Code setup I use to build all of this — the memory system, the commands, the workflow. If you want to see how the sausage gets made, it's all there.
The bet
The bet is simple: a decade of building software in public creates more leverage than a decade of doing excellent work in private where no one can see it. I might be wrong. I'd rather find out in the open.
If you want to follow along, subscribe below. Next dispatches will go deeper on individual builds — what worked, what broke, and what I'm shipping next.