Building an AWS solutions architect agent with Amazon Bedrock
summary
Builds an AWS solutions-architect agent on Amazon Bedrock that combines a Bedrock knowledge base (S3 → vector store/OpenSearch serverless) with Lambdas as actions to answer questions from docs and generate Terraform code saved to S3, then runs an infra-cost tool (infraCost in Docker via Lambda) to estimate infra cost. Demo syncs 10,032 files into the KB, produces Terraform artifacts and cost estimates (~$270–$600), and shows model-choice tradeoffs (up to ~60× or 12× cost differences between models) with Bedrock tracing for decision visibility.
AI-written summary · about
chapters (12)
- 0:00Introduction and speakers
- 3:34Why automate solutions architect tasks
- 7:50Components of a generative AI agent
- 10:21Building a knowledge base tool
- 15:42Answering questions with knowledge base
- 17:05Choosing the right AI models
- 20:13Generating infrastructure code
- 23:05Estimating infrastructure costs
- 27:18Combining tools into an agent
- 30:50Configuring the Bedrock agent
- 33:20Testing the agent with a demo
- 36:13Key takeaways and next steps
related talks
29mBuild and scale reliable multi-agent systems on AWS
43mBeyond Chatbots: How Amazon Bedrock Agents Drive Enterprise AI
31mLamRAG: From Data to Constructive Insights using Amazon Bedrock
23mBuilding Production Multi-Agent RAG Systems on Serverless AWS
28mBuilding Agentic AI applications on AWS
24mBuilding GenAI applications with Amazon Bedrock
39mGenerative AI Solution with AWS Serverless and Bedrock
33mShipping Enterprise AI Agents with Amazon Bedrock AgentCore