Craft your own Generative AI Chatbot with Amazon Bedrock
summary
Demonstrates building a Slack generative chatbot called Assistant that runs on Amazon Bedrock, uses RAG with embeddings stored in an open source vector store, and orchestrates updates with Airflow into Snowflake and S3. Code uses LangChain (or LlamaIndex) for retrieval and prompts, API Gateway exposes the bot, CloudWatch monitors usage, and guardrails filter sensitive content. Covers chunking strategies, multi‑index tradeoffs, max token limits and a two‑step turbo→regular model fallback to reduce hallucinations.
AI-written summary · about
chapters (5)
- 0:00Introduction and agenda
- 8:03What is Amazon Bedrock
- 12:58Building a chatbot with Bedrock
- 25:38Best practices and tips
- 36:12Conclusion and resources
related talks
49mBuilding AI Agents with Spring & MCP
43mBeyond Chatbots: How Amazon Bedrock Agents Drive Enterprise AI
47mOutsmarting the Genie: Smart Architecture Decisions for GenAI Apps
20mBuilding a secure end-to-end generative AI application in the cloud
44mBuilding a Domain Specific GenAI Chatbot with Serverless
36mGENAI-Powered chatbot to increase team efficiency
28mBuilding Agentic AI applications on AWS
27mBuilding AI Agents with Spring & MCP