01
Found the real constraint
I was hired to run marketing. The harder problem was operational: important work lived across inboxes, spreadsheets, vendor portals, and systems that could not talk to one another.
About / Spencer D'Avis
I learned engineering by becoming responsible for the result.
No CS degree. No inherited codebase or development team. I started with an operator's view of the business, then built the software that the business needed.
02
Built the first system
I taught myself PHP and wrote a 14,000-line maintenance application before I had used version control. It proved the workflow and became the foundation for the production platform that followed.
03
Rebuilt for production
I founded the successor web and mobile projects, then architected the rewrite and directed contract developers through later build-out. The platform now coordinates maintenance work across web, iOS, and Yardi.
04
Made verification part of the product
The work moved into applied AI: receipt extraction, classification, virtual staging, and model evaluation. I learned that the valuable question is rarely whether a model can produce an output. It is whether the system knows when that output is wrong.
How I work
Business context is part of the architecture.
I work best where the requirements are incomplete and the workflow crosses organizational boundaries. I can sit with the people doing the work, find the actual constraint, and carry the solution through the database, interface, integration, rollout, and production support.
I prefer explicit failure modes over confident guesses. That shows up in human verification links, classifiers that return null when evidence conflicts, regression tests written after incidents, and deterministic checks around generative models.
Public evidence