What we have learned building and running these systems - written for the engineers who will maintain them.
July 28, 2026
Remote state, locking, and workspace layout decisions that stop being cosmetic the moment a second engineer runs an apply.
ReadJune 15, 2026
Why compliance sprints keep recurring, and how to design controls so the evidence produces itself.
ReadHow a logistics platform moved from hand-managed EC2 to Kubernetes incrementally, keeping deploys running throughout.
Read the case studyMay 9, 2026 · 2 min read
What changed
Deploy time cut from ~40 minutes to under 6
Environment rebuild from code in ~25 minutes
Zero customer-facing downtime during cutover
Building a retrieval-augmented assistant for a fintech support team, and the evaluation work that made it trustworthy.
Read the case studyMarch 22, 2026 · 2 min read
What changed
Median first-response time down from 4h to 35m
Deflection on 31% of tier-1 tickets
Abstains rather than answering on out-of-scope questions