How I moved briefcast from a local AI briefing app into a production-style AWS setup with queues, workers, storage, observability, and cost trade-offs.