Hands-on project · AWS · AI-assisted
AI-Enabled DevOps in AWS
Build, secure, and operate a production-style deployment on AWS using modern automation — while practicing AI-assisted workflows across the full software delivery lifecycle.
Who this is for
Software engineers who want real AWS delivery experience.
Junior-to-mid DevOps practitioners and backend developers who want hands-on experience deploying a real application to AWS with modern automation and AI-assisted tooling — not a cert cram or slide deck.
Prerequisites
- →Comfortable using Git for everyday work — clone, branch, commit, push, open PRs.
- →Ability to read and run a Java/Spring Boot codebase at a basic level — deep Spring expertise not required.
- →Laptop capable of running Docker Desktop and AWS credentials in the standard local chain (
~/.aws/config). - →No prior AWS experience required — the course starts from account setup.
- →Willingness to use AI coding assistants as tools under human review — not as a source of truth for production secrets or unverified infrastructure.
What you'll build
Pretzel Shop — from ./gradlew bootRun to ECS Fargate.
A full-stack pretzel ordering app: Spring Boot 3.x API (Java 21, Gradle) with React/Vite frontend, PostgreSQL for persistence, Redis for caching. Over 12 sections you take it from your laptop to a fully automated AWS deployment.
Local stack
Docker Compose with Postgres, Redis, and the API. Secrets in .env — never in Git.
AWS infrastructure
VPC across AZs, RDS PostgreSQL, ElastiCache Redis, ECR, ECS Fargate, Application Load Balancer — all provisioned with Terraform.
CI/CD pipeline
GitHub Actions: build, Trivy scan, push to ECR tagged with commit SHA, deploy to ECS — authenticated via GitHub OIDC, no long-lived keys.
Operations & security
CloudWatch dashboards and alarms, structured logging, autoscaling, RDS backup drills, OWASP Dependency-Check, Checkov, IAM Access Analyzer, and incident runbooks.
You also build an in-repo developer platform — AGENTS.md, .cursor/rules/, playbooks, Makefile targets, and a PLATFORM-MAP.md that indexes the entire toolchain.
Roadmap
12 sections in four parts.
Foundations
DevOps mindset, Git, AI-assisted development, AWS accounts, IAM, CLI, and cost guardrails.
Containers & CI
Multi-stage Dockerfiles, Docker Compose local stack, GitHub Actions CI with test gates and artifact upload.
IaC & Deploy
Terraform for VPC, RDS, ElastiCache, ECR, ECS, ALB — then CI/CD to staging and production with Secrets Manager.
Ops & Capstone
Observability, security gates, day-two operations, and a full SDLC capstone with cost review and responsible AI norms.
Tech stack
Tools you'll use end to end.
Learning outcomes
What you'll be able to do.
- ✓Provision AWS networking and managed services (VPC, RDS, ElastiCache, ECR, ECS, ALB) with Terraform using
fmt → validate → plan → apply. - ✓Configure GitHub Actions to build, test, scan, and deploy container images without long-lived credentials — GitHub OIDC to STS.
- ✓Store and retrieve secrets through AWS Secrets Manager and SSM Parameter Store, with ECS task definitions referencing ARNs instead of plaintext.
- ✓Build CloudWatch dashboards and metric alarms for ALB latency, HTTP 5xx errors, and ECS resource utilization.
- ✓Wire Trivy, Checkov, and OWASP Dependency-Check as CI gates that block merges on critical findings.
- ✓Configure ECS autoscaling with target-tracking policies and deployment circuit breakers for automatic rollback.
- ✓Execute an RDS backup and restore drill, and write operational runbooks for common ECS failure modes.
- ✓Use AI assistants (Cursor with MCP servers) to draft Terraform, workflow YAML, and tests — while maintaining human ownership of every merge, apply, and deploy decision.
Ready to ship on AWS?
Book a call to talk through fit, how the project works inside Senior Ready, and whether this track matches where you are today.
Book a call →How AI fits in
Cursor + AWS MCP as pair-programming — you own every merge and apply.
Throughout the track, Cursor with AWS MCP servers drafts Terraform, GitHub Actions YAML, tests, log queries, and threat models. AI is treated like a junior pair: useful for first drafts and spotting patterns, never a substitute for reading diffs, running terraform plan, or reviewing IAM trust conditions yourself. Every merge, apply, and deploy decision stays with you.