Insights / Topics
The engineering topics we actually write about — each one earned in the work.
01 / Topics
Retrieval-augmented generation that survives real users and real data.
Building on Bedrock for production workloads — model choice, quotas, guardrails, and cost.
Observability, evaluation, and the operational plumbing that keeps AI systems honest in production.
Measuring whether an AI system actually works — before you trust it, and after.
Keeping AI and cloud bills from quietly becoming the incident.
Deciding whether to build at all — and, if so, what to build first.
Scoping AI pilots so they tell you the truth about production readiness.