About

I'm a software architect. I've spent most of my career on software that handles other people's money: accounts payable and procurement today, and before that, fourteen years building mortgage lending software.

These days that mostly means working out where an AI model belongs in that kind of system, and where it doesn't. The rule I keep coming back to: the model can assess and explain, but a person decides. A grade with no reason a person can read isn't worth much, and a confidence score is only useful if you've checked it means something.

The thread through all of it is the same problem in different clothes: many messy outside sources, normalized into one clean internal model. Credit vendors, then ERP systems, now invoices and receipts read by AI.

I write here about what I learn doing that, and about setting up AI coding tools so they help more than they guess. I'm in Arizona. The best way to reach me is LinkedIn.