1 September 2026
AI as a production concern
I am adding LLM skills the same way I ship backends: auth, failure modes, and maintainability — not notebook demos.
My job is PHP/Symfony on live systems. I am learning how LLMs can sit beside those systems — retrieval, evaluation, authentication, timeouts — with the same bar as a billing job or a patient-portal API.
I do not list a public RAG product. The value I am building toward is operational: a model-backed feature an organisation can run, observe and change.
