Tutorial: From Planni...
summary
Hands-on steps deploy a production-ready RAG chatbot atop the OPEA blueprints using a provided EKS cluster (workshop studio access ~3 hours). Architecture composes microservices (embeddings, retriever, reranker, “mega” orchestrator) behind an engine X gateway and a vector DB called radius; the inference pod pulls a TGI 8B (~30 GB) model. Workflow shows indexing a Nike PDF into vectors, retrieving top‑4 docs then reranking (scores ~0.99, 0.80, 0.53, 0.04), and outlines upgrading retrieval via graph RAG with Neo4j to connect chunks, summaries and structured data.
AI-written summary · about
chapters (6)
- 0:00Open Platform for Enterprise AI
- 9:17Workshop setup and AWS access
- 24:30Deploying RAG with OPA
- 1:00:07RAG workflow and embeddings
- 1:10:15Graph RAG benefits
- 1:30:18Q&A on Graph RAG
related talks
40mRolling out Enterprise AI: Tools, Insights, & Team Empowerment
44mThe RAG Really Ties the App Together
41mFrom RAGs to riches: Build an AI document interrogation app in 30 mins
45mLessons learned in bringing a RAG chatbot with access to 50k+ diverse documents to production
25mComposable RAG Pipelines With OPEA Highlighting GraphRAG
25mSet Up an Open Source RAG in 30 Minutes
59mRAG Against the Machine
26mWelcome to the AI Jungle! Now What?