Case studies
What the problem was, what I built, and what changed. Short reads, a few minutes each.
An LLM turns plain-English queries into structured Elasticsearch queries. A semantic cache and hybrid retrieval sit behind it, and a RAG layer answers.