Beyond Keywords: Building an AI-Powered Search That Reads Your Mind
summary
Shows how an OpenSearch-based pipeline converts short queries into embeddings and expanded JSON filters using Model Context Protocol plus a Strands agent that calls an LLM (demo uses a CLAW model) to produce persona-aware queries. Demonstrates hybrid search (BM25 + vector embeddings), ML inference processors for indexing and re-ranking, semantic caching to avoid repeat LLM calls, and a fallback path with 100–200 ms baseline and a 500 ms LLM timeout for production SLAs.
AI-written summary · about
chapters (8)
- 0:00AI-powered search intro
- 3:10Keyword search challenges
- 7:24Search technology evolution
- 10:49Intent in search
- 16:25Personalized search architecture
- 20:48Live demo personas
- 27:07AI search production tips
- 29:13Resources and closing
related talks
51mQuery Understanding With Large Language Models: Techniques and...
22mAgentic Hybrid Retrieval: Orchestrating LLM Agents, Hybrid Search & Explainability in OpenSearch
38mAI-Ready Data Foundation Using Amazon OpenSearch
38mNatural Language Search-Lexical and Semantic
39mExperiments with OpenSearch and AI
47mBoosting AI With OpenSearch: Vector Stores, Similarity Search, and Retrieval-Augme...
21mBuilding RAG With OpenSearch ML Plugins: From PDFs To Voice-Enabled Search
26mOpenSearch: Conversations with Context