Notes from inside the engagements.
What we're learning about engineering, delivery, product, and design, written by the people doing the work.
Why we still write architecture decision records for two-week projects
The habit looks like overhead until the second engineer joins a project six months later and has to guess why a decision was made. Here's the format we've settled on after a decade of getting it wrong.
Read the pieceLatest
Jun 2026The migration checklist we run before touching production data
Every step we take before a schema change goes anywhere near a live database.
What a discovery week with Kestrel actually looks like
Day by day, from the first working session to the scope document you'll sign off on.
Designing for the claims adjuster who's had four cups of coffee
What the Palisade Insurance redesign taught us about interfaces built for tired experts.
More reading
Mar 2026Postmortem: the Friday deploy we shouldn't have shipped
A rollback, a root cause, and the two process changes that came out of it.
How we scope fixed-price work without guessing
The estimation method we use, and the two questions that catch most bad scopes early.
A defense of boring technology
Why we've shipped six figures worth of production systems on tools nobody writes conference talks about.
Onboarding a new engineer into someone else's codebase
Our checklist for the first two weeks on an inherited system, before touching anything critical.
Have a project we should be writing about instead?
Tell us what you're building and let's see if it's a fit.