Intelligent IaC Generator Agent

In this project, you'll build CloudForge AI

• an agentic infrastructure-as-code generator for a DevOps consultancy

that needs to accelerate cloud deployments by converting plain English architecture descriptions into production-ready CloudFormation templates. What makes this system truly agentic is the self-correction loop: when the AI generates a template that fails validation, it automatically feeds the errors back to Amazon Bedrock and asks it to fix the issues — repeating up to 3 times until the template passes. This autonomous iteration is the core pattern behind modern AI agents. You'll learn to build agentic AI workflows on AWS: • Amazon Bedrock for AI-powered CloudFormation template generation from natural language • AWS Lambda functions for generation, validation, refinement, and template management • AWS Step Functions for orchestrating the agentic loop with Choice states and retry logic • Amazon DynamoDB for tracking generation requests, retry counts, and audit logs • Amazon S3 for template storage with versioning across drafts, refined, and validated stages • Amazon CloudWatch for monitoring, logging, and error alerting • Amazon API Gateway for exposing the generation pipeline as a REST API • AWS IAM for least-privilege security across all services This project showcases the most important pattern in agentic AI — autonomous self-correction — on AWS, teaching you how to build intelligent automation that improves its own output through iterative feedback loops.