Vectors Vs. Hallucinations: OpenSearch's GenAI Survival Kit
summary
Shows concrete RAG pipelines that reduce hallucinations by 80%+ error reduction using OpenSearch vector storage, the k‑NN plugin and hybrid BM25+vector fusion, with examples of ingesting embeddings at scale and achieving sub-second k‑NN queries. Demonstrates chunking strategies (fixed, sentence window, parent–child), RAG over petabyte traces, eval metrics (precision, NDCG), prompt+ranker interactions, and Kubernetes deployment blueprints plus code snippets and failure autopsies for reproducible production rollouts.
AI-written summary · about
chapters (8)
- 0:00Hallucinations in GenAI
- 3:32Fine-tuning vs RAG
- 6:39RAG in OpenSearch
- 9:33Hybrid search methods
- 11:38Chunking strategies
- 14:05Testing RAG pipelines
- 16:52Prompt engineering with RAG
- 17:15Summary
related talks
28mUse OpenSearch as Your Vector Search Database To Optimize GenAI Accuracy!
47mBoosting AI With OpenSearch: Vector Stores, Similarity Search, and Retrieval-Augme...
41mGenAI, vector/semantic/hybrid search, RRF, NLP, (...) and other buzzwords
33mVector Databases in the Age of AI: Embracing RAG for Enhanced Search Capabilities
20mChallenges in Building RAG Systems (and Solutions Using OpenSearch)
48mEnter the Brave New World of GenAI with Vector Search
37mBuilding advanced RAGs with Elasticsearch OpenSearch
45mMaking semantic search & RAG real: How to build a prod-ready app