
AISample
Evaluation first: taking AI features from demo to production
Most AI prototypes fail in production for the same reason: nobody can measure quality. Here is how we build the measurement before the feature.
· 6 min read
Practical writing on AI, architecture and design, from the people doing the work.

Most AI prototypes fail in production for the same reason: nobody can measure quality. Here is how we build the measurement before the feature.
· 6 min read

One page per decision, stored next to the code. Why we write architecture decision records on every engagement, and what goes in them.
· 5 min read

If a component exists only in a design file, it is a suggestion. How we build design systems that survive contact with engineering.
· 5 min read