Field notes on engineering for the AI age.
Retrieval that looks great on a slide often falls apart on real documents. Here's the evaluation loop we use before anything ships.
The questions we ask before writing a line of code — and the ones we deliberately skip.
When an off-the-shelf model is enough, and when you actually need to own the stack.
How to turn an AI-generated prototype into software real customers can trust.
What the Model Context Protocol is, why it matters, and where it fits.
Compliance without hiring a compliance department — what to do first.
Why automations fail adoption, and how to design ones that stick.