We write about the patterns we keep meeting in production: AI system design, RAG reliability, offline sync, automation architecture, and the commercial side of shipping software.
Off-the-shelf retrieval augmented generation pipelines optimise for demos. Production is different.
Your point-of-sale system will lose connectivity. Here is how we handle it without losing a single transaction.
LLMs are bad at arithmetic. Here is the pattern we use to make our agents reliable with financial data.
How we keep one design system in sync across a React web app and a React Native mobile app.
Building an ordering system on WhatsApp Business API ÔÇö what works, what does not, and what we would do differently.
Four metrics we track in production to know if our agents are actually working.
Scope creep does not happen because clients are difficult. It happens because projects are priced incorrectly.