Notes from
production.
Field reports, decision records, and the occasional rant. Written by engineers, for engineers and the founders who hire them.
AWS vs Azure vs Google Cloud: A 2026 Comparison for Choosing Your Cloud Provider
A practical 2026 comparison of AWS, Microsoft Azure, and Google Cloud Platform — pricing, compute, databases, AI/ML, and DevOps — with a decision framework to pick the right cloud provider for your workload.
Stop calling it RAG. Start calling it search that explains itself.
Why hybrid retrieval beats vector-only in production — and the eval suite we ship with every implementation. With code, latency numbers, and the prompt template we keep coming back to.
From Docker to a real on-call rotation in eight weeks.
The pragmatic infrastructure ladder we use for series-A startups, with cost numbers and tooling choices.
Multi-tenant or dedicated? A decision tree for B2B SaaS.
How we choose between shared, isolated, and dedicated tenancy — with billing, ops, and risk in the same frame.
Tool-using agents that don't go off the rails.
Designing the schema, the retry policy, and the human-in-the-loop checkpoints. With three production examples.
Postgres as a queue: when it's fine, and when it isn't.
SKIP LOCKED has come a long way. Where it works, where it breaks, and the migration path to a real broker.
React Native in 2026 — when it's the right call.
The questions we ask before recommending RN. App-store reality check, hiring math, and the cases where native still wins.
Why we write architecture decision records — even for two-week projects.
Five-minute writeups that save five-hour debates six months later. Our template, and the rules for when to write one.