Infrastructure as Code with Terraform

Chapter 3 of 6 — AI Cloud Engineer Roadmap

Everything you built by hand in Chapters 1–2 — the VPC, the IAM roles, the EKS cluster — gets rebuilt in this chapter as Terraform: declarative, versioned, reviewable infrastructure instead of a trail of console clicks nobody can audit six months later.

What you'll build: the VPC and EKS setup from Chapters 1–2, re-expressed as Terraform modules — the first half of a full IaC pipeline that Chapter 4 finishes by wiring it into CI/CD.

Tools: Terraform

Where AI helps: AI scaffolds the module structure and boilerplate resource blocks quickly — you still own state management (remote state, locking, workspaces) and the decisions about what's a module boundary versus what's inline. State file mistakes are the most common way Terraform sessions go sideways, and no amount of AI-generated HCL fixes a corrupted or conflicting state.

Modules in this chapter

• Terraform Basics with AWS — providers, resources, variables, state, your first terraform apply

Why this matters

The difference between "I built a VPC" and "I can rebuild this VPC identically in a new account in five minutes" is Terraform. Once your infrastructure is code, it can be reviewed in a pull request, diffed before it's applied, and — critically for Chapter 4 — triggered automatically by a CI/CD pipeline instead of run by hand from your laptop.

---

Next: CI/CD & Delivery Pipelines

Chapter 4 takes the Terraform from this chapter and puts it behind GitHub Actions: every push triggers a plan, every merge triggers an apply, with security scanning gating the pipeline before infrastructure changes reach AWS.