From Semantic Search to AI Agents: Quick Start on Production-Ready ML Appl...
summary
opensearch-ml-quickstart provides end-to-end automation for connector creation, model registration, ingest pipelines and search pipelines, turning weeks of setup into scripted commands; the repo contains about 10 examples and demos show lexical, dense and sparse embeddings, and hybrid search working together. Connectors talk to SageMaker, Bedrock, OpenAI or local models and register embeddings (Barro Titan) and LLMs (Barro Clonet 4); a conversational agent uses those tools, runs multi-step retrieval, returns summaries and recommendations and even finds ~400 diffuser results in the demo.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 0:22Lexical to agentic search
- 3:14OpenSearch in AI agents
- 5:22Models with OpenSearch connectors
- 7:29Hybrid search pipelines
- 10:16Motivation for ML Quick Start
- 11:19Abstraction layers in ML Quick Start
- 13:27Local and AWS deployments
- 14:08CLI and wrappers setup
- 15:34Workflow templates overview
- 16:15Demo and repo
- 18:47Lexical search demo
- 21:46Semantic search with embeddings
- 24:47Hybrid search demo
- 26:23Automating with templates
- 29:08Agentic search demo
- 33:41Summary
- 34:03Q&A
related talks
43mOpenSearch Flow and Agent Frameworks: Easy AI
38mA Visual AI Search Flow Builder for OpenSearch: Build GenAI and Vector...
36mBuilding Next-gen Search With OpenSearch
30mBeyond Keywords: Building an AI-Powered Search That Reads Your Mind
38mAI-Ready Data Foundation Using Amazon OpenSearch
47mBoosting AI With OpenSearch: Vector Stores, Similarity Search, and Retrieval-Augme...
40mBuilding Intelligent Search Systems With Agentic RAG and OpenSearch
37mAgentic AI Meets Search: Real-World Use Cases With...