Credit Workbench
I wanted to see exactly where a levered credit breaks — so I built an engine that computes the downside and lets you stress it live.
Understanding comes first. Software is simply the medium.
I wanted to see exactly where a levered credit breaks — so I built an engine that computes the downside and lets you stress it live.
A data investigation into the gap between reported and economic health across private-credit portfolios — built from public filings, industry-level only.
A written investment memo on a single name — the piece where the judgment does the work and the model just supports it.
New explorations land here as I chase them down. If you're working on an interesting financial problem, that's often where the next one starts.
Start a conversationIf I can't explain the number, I don't use it.
Understanding today beats a forecast I can't defend.
A model is only as honest as the assumptions behind it.
AI does the work; the judgment stays mine.
I don't build models to predict the future. I build them to understand what breaks first.
I work in investment research and private markets, and I enjoy the intersection of finance, software and analytical thinking.
Whenever I hit a financial problem that feels difficult to reason about, I tend to end up building something to understand it — sometimes a model, sometimes code, sometimes both. The goal is never the software. It's the understanding.