
I’ve spent a long time focused on automation and reliability. These days I’m publishing new work, sharing notes as I go, and gradually turning old ideas into real projects.
This site is a living log of what I’m building. LinkedIn is my resume.
Highlights
Supports multiple sites with consistent structure and rules
Clear draft vs published behavior
Theming designed to work in light and dark mode
I’ve spent most of my career working on the kinds of systems where reliability matters, the work sits behind the scenes, and small mistakes can become big problems. That shaped how I think, I prefer clear inputs, predictable outputs, and tooling that makes work feel calmer.
What I’m optimizing for now
At this point, I’m most interested in roles that feel focused and sustainable. Smaller teams, clearer priorities, and a pace that leaves room for deep work. I’m remote-first, but flexible when the role fits.
I’m not trying to “escape” any industry. I’m choosing the kind of environment and cadence that helps me do my best work. If the right opportunity shows up in finance or another high-compliance space, I’m open to it.
What I’m doing right now
I’m working on Site Engine, a foundation for running multiple sites from one engine. It lets me ship without rebuilding basics each time, and it gives me a place to explore practical AI-assisted workflows, in a way that stays grounded in real constraints.
How I like to work
Clear rules and clean data shapes
Small slices, shipped and verified
Observability and failure modes considered early
Simple UI, strong backend discipline
If you’re hiring
I’m a strong fit for backend engineering, data-heavy work (SQL, ETL, pipelines), automation, and internal tools. I can’t share past employer code, but I can talk clearly about architecture, tradeoffs, and how to make systems more reliable over time.
If you’re a builder
I keep notes on decisions, mistakes, and what actually worked while building Site Engine and other experiments. No hype, just the process.

My first reader marks up drafts in pencil. Getting her notes into the editor meant anchoring every mark to the words it touched, never to a page or a position.

An invite link failed three seconds before my friend tapped it. Hotmail's link scanner had already spent the one-time token. The fix was a button.

Five times my health app's chat said a number was in my records but out of its sight. I stopped widening the window and gave it a way to ask for rows.

I reproduced a published benchmark exactly: same accuracy, same manifest hash, same token count. Then the confusion matrix showed the headline was still misleading.

I built a counter to measure four years of my own chat history, got a clean 41%, and then found out the number was almost entirely a proxy for turn length.

A week of pulling museum accession numbers and biblical texts taught me that a 403 is honest and a 200 can be a lie. The fix was to stop being a client and open a browser.

I wrote a non-negotiable rule about never making a web request wait on AI, then broke it in my own health feature. Three bugs later I saw they were all the same bug.

I pulled 800 readings off a blood glucose meter with a browser tab. The meter was right about all of them. My parser was six hours wrong about every single one.

Five separate bugs in one week of training-data work, and every one of them was the same bug: I had been auditing the description of the data instead of the data.

A single commercial registered-agent suite on Kelly Johnson Blvd is the principal address on 4,495 El Paso County entities. Strip it out and the monthly formation series goes completely flat. No wrongdoing, just a measurement problem hiding in a public statistic.

Between 2014 and 2025, five private sectors holding 76,332 El Paso County jobs saw their average real wage stall or fall. Health care, the county's largest private employer, gained $23 a week in eleven years.

One in five new Colorado Springs jobs over the last decade was in health care, and for the last two years it is the only part of the market still growing. A QCEW data brief on the sector carrying the county, and the paychecks that stood still.

An economic brief I published rests on a single number. The query was six lines. The hard part was deciding what "closed" and "in the county" actually mean across 1.5 million messy public records.

For a decade, Colorado Springs started about four businesses for every one that closed. In 2024 that ratio fell to 2.53 to 1, its lowest on record. Here is what the Secretary of State filings show, and why the closures landed where they did.
Have a question or want to work together? Fill out the form below and I'll get back to you soon.