Serverless & Agentic AI: Better Together
summary
Claims that agents are inherently event-driven and map cleanly to AWS serverless: API Gateway → Lambda agents calling Bedrock or SageMaker for LLM inference, tools exposed via an MCP server (which can run on Lambda), and DynamoDB used for chat memory while Aurora/Postgres, OpenSearch or S3 vector search hold embeddings. Shows a working pattern: Lambda agent calls LLM, evaluates output in a separate function and re-invokes the agentic loop until satisfactory; Step Functions and EventBridge orchestrate swarms and parallel tasks.
AI-written summary · about
chapters (10)
- 0:00Intro
- 0:40What is an agent?
- 4:26Events
- 5:48LLM inference
- 6:29Agents/Tools/MCP
- 7:33Data
- 8:46Agentic RAG
- 10:06Swarm
- 11:11Agent orchestration
- 14:10Outro
related talks
28mBuilding Agentic AI applications on AWS
31mAgentic AI architecture patterns
41mServerless LLM with AWS Lambda and LangChain: A Revolution in Application Development
45mServerless & Event-driven Patterns for GenAI
23mBuilding Production Multi-Agent RAG Systems on Serverless AWS
44mBuilding a Domain Specific GenAI Chatbot with Serverless
17mTransforming Robotics Service with Agentic AI on
48mHow to Make Smart Architecture Decisions when Building Gen AI Apps