LamRAG: From Data to Constructive Insights using Amazon Bedrock
summary
An internal app called Feedback Key centralizes ~500 sprint feedbacks and uses Amazon Bedrock playground for prompt tuning, then creates a knowledge base backed by a vector database (OpenSearch serverless) with Titan embeddings (1,536 dims). Default chunking fails so files use no chunking to keep per‑user context. Agents call AWS Lambda to enforce admin-only queries, and choreography uses prompt flows to split classification, retrieval and synthesis. The pipeline is built and reproducible and the prototype takes under 30 minutes to assemble.
AI-written summary · about
chapters (8)
- 0:00Introduction and Session Overview
- 0:59Challenges in Feedback Collection
- 2:13Implementing AI for Feedback Analysis
- 6:27Introduction to Vector Databases
- 8:10Creating a Knowledge Base with Amazon Bedrock
- 16:39Advanced Query Handling with Agents
- 21:38Future Plans and Ambitious Goals
- 29:25Conclusion and Final Thoughts
related talks
41mServerless LLM with AWS Lambda and LangChain: A Revolution in Application Development
28mBuilding database queries with Amazon Bedrock
38mGenerative AI - Architectures and applications in depth
24mBuilding GenAI applications with Amazon Bedrock
20mData is Your Differentiator: Building Secure and Tailored AI Systems
55mSecurely build generative AI apps & control data with Amazon Bedrock
40mBuilding an AWS solutions architect agent with Amazon Bedrock
29mCustomizing genAI applications for your business