
Set up your cloud development environment by installing node.js, configuring aws cli, and provisioning infrastructure with terraform; prepare next.js app and domain via route 53, with github runners and cursor.
Plan phase guides scaffolding a Next.js TypeScript image upload gallery with a private S3 bucket and signed URLs, plus Terraform and GitHub Actions to ECS Fargate.
Configure a github actions ci/cd workflow with oidc authentication, build and push a next.js app image to ecr, and deploy to ecs on main with health checks.
Review the end-to-end journey from a blank folder to a live domain app with automated redeploys via GitHub, and learn to clean up AWS resources and Terraform destroy.
What does it actually look like to go from an empty folder to a real application running in the cloud while letting an AI coding agent do the coding?
This hands-on workshop takes a different approach from a traditional course.
Instead of learning one technology in isolation, we'll bring multiple tools together and use an AI coding agent inside Cursor to take a project through the entire journey from code to cloud.
And there's another important difference: we won't hand-write the application or infrastructure code ourselves.
We'll start with an empty folder and direct an AI agent to scaffold a Next.js application. From there, we'll containerize it with Docker, define its AWS infrastructure using Terraform, push everything to GitHub, and create a GitHub Actions CI/CD pipeline that automatically builds and deploys the application.
Finally, we'll connect our own domain and configure HTTPS so that our application is accessible through a real URL with a valid certificate.
Throughout the workshop, we'll work with:
AI-assisted development with Cursor
Next.js
Docker and containerization
Terraform and Infrastructure as Code
AWS cloud infrastructure
GitHub
GitHub Actions
CI/CD
DNS and custom domains
HTTPS and certificates
Automated application deployments
But this workshop isn't really about watching an AI agent generate code.
The important part is learning how to direct it.
You'll see how to describe requirements, review generated code and infrastructure, verify what the agent is doing, and decide whether its proposed approach actually makes sense.
And things will go wrong.
We'll encounter real issues and instead of editing everything manually or hiding those problems, we'll work through them with the AI agent.
That's an important part of the workshop.
AI coding agents are non-deterministic. They can make different decisions, misunderstand requirements, and generate solutions that don't work on the first attempt. Learning how to recognize those problems, provide additional context, and guide the agent toward a working solution is part of modern AI-assisted development.
By the end of the workshop, we'll have taken a project from an empty folder to a complete cloud deployment with an automated CI/CD pipeline, custom domain, and HTTPS.
More importantly, you'll have seen what it actually looks like to use an AI coding agent across application development, containers, Infrastructure as Code, CI/CD, and cloud infrastructure while remaining in control of the process.
This workshop is a great fit if you are:
A developer exploring AI-assisted software development
A cloud or DevOps engineer interested in coding agents
A full-stack developer wanting to connect development and infrastructure
Someone already familiar with individual cloud and DevOps tools who wants to see how they work together
Anyone interested in using AI agents for more than generating isolated pieces of code
This workshop assumes basic familiarity with technologies such as Docker, Terraform, AWS, GitHub, and web development. We won't teach each technology from scratch. Instead, we'll focus on combining them into one complete workflow.
If you're looking for individual technology fundamentals, my Hands-On Introduction courses cover those topics separately.
If you're ready to put the pieces together and see how far an AI coding agent can take a real project, let's start with an empty folder and move from code to cloud.