AI ops

A practical playbook to manage LLM costs in production: choose models wisely, cut tokens, batch and cache calls, use retrieval over generation, and estimate spend.