Notes on building production AI — the engineering around the model.
How a manual grantee review measured in hours became a 15-minute workflow — and what I learned shipping an AI system that a compliance team trusts with real decisions.
Read post →“The hard part of AI engineering is rarely the model. It's everything around the model — data protection, cost discipline, grounding, and earning trust.”
A recurring headache in accounting, a two-day company hackathon, and a small team that decided to fix the problem instead of just complaining about it. Now in production.
Read post →“The idea was not 'let's build an AI tool.' It was 'we keep catching the same mistakes too late. What if we caught them before the invoice ever left the building.'”
We wrote perfect documentation for a new system, then accepted that nobody reads documentation. So we turned it into something you can talk to, with a human always one escalation away.
Read post →“Documentation does not fail because it is badly written. It fails because it sits somewhere people do not look. The best user guide is the one that answers back.”