Learn Claude Code.
Deploy Real Infrastructure.

A hands-on course for DevOps, SRE, QA, and developers. Go from your first claude command to a live app on Kubernetes — with AI as your co-pilot the entire way.

14 Blocks
1 Real App
1 Live Deployment
Start Course → Get Mentoring

What You'll Build

Not a toy demo. A real Next.js dashboard app, containerized, deployed to k3s on DigitalOcean via GitOps, accessible from anywhere on the internet.

Before

ai-coderrank: a coding AI models leaderboard.

Running locally. Light theme. No infrastructure.

npm run dev → localhost:3000

After

Dark theme. Docker image. CI/CD pipeline. k3s cluster. ArgoCD GitOps. Public IP.

http://<droplet-ip>:30080 → live on the internet

Claude Code was used to build parts of Claude Code itself. In this course, you'll use Claude Code to deploy an app that ranks AI coding models — including Claude. It's AI all the way down.

Who This Is For

If you work in tech and haven't tried Claude Code yet, this is your on-ramp.

>_

DevOps Engineers

Automate infrastructure tasks, write IaC, debug pipelines, and manage deployments through conversation.

SRE

Analyze logs, create runbooks, build monitoring skills, and set up automated health checks with Claude.

QA Engineers

Generate tests, review code for edge cases, automate test workflows, and build quality gates with hooks.

{ }

Developers

Ship features faster, get instant code reviews, manage git workflows, and understand unfamiliar codebases.

AI-Curious Practitioners

You use a terminal daily but haven't tried AI tooling yet. This is your practical on-ramp — real tasks, not toy demos.

Why Claude Code in 2026

Claude Code isn't just autocomplete. It's a full-session AI agent that reads your codebase, runs commands, edits files, manages git, and orchestrates multi-agent workflows. Here's what you'll master:

Plans & ADRs

Claude thinks before it acts. Enter plan mode, generate Architecture Decision Records, and create Mermaid diagrams — all from a conversation.

Custom Skills

Package team workflows as reusable slash commands: /review-k8s, /check-docker, /deploy. Your playbook, Claude-powered.

Hooks & Automation

Auto-format on save, block destructive commands, notify on completion. Hooks run shell commands at every lifecycle point.

MCP Servers

Connect Claude to GitHub, Slack, Linear, databases, and 50+ external tools via the Model Context Protocol.

Sub-agents & Teams

Spawn specialized AI workers for security review, K8s validation, or parallel research. Coordinate multi-agent teams on complex tasks.

CI/CD & GitHub Actions

Tag @claude in a PR for automated review. Create issues and let Claude implement them. Schedule recurring cloud tasks.

The Course

14 blocks. Each one builds on the last. By the end, you've deployed a real app and mastered Claude Code.

00

Welcome & Setup

Install Claude Code, authenticate, fork the project, and have your first AI conversation.

install /help
01

Understanding Your Codebase

Let Claude analyze the project, generate CLAUDE.md, and explore the architecture.

/init CLAUDE.md
02

Running & Testing

Set up locally, run tests, build Docker — and watch Claude handle errors.

Bash tool permissions
03

Planning with ADRs & Diagrams

Think before you code. Create ADRs, Mermaid diagrams, and an implementation plan.

plan mode ADR Mermaid
04

Making Changes — Dark Theme

Implement the dark theme, preview it, and commit — Claude handles the git workflow.

Edit/Write git
05

Memory & Project Intelligence

Teach Claude your preferences, team rules, and project conventions — across sessions.

.claude/rules /memory auto memory
06

Custom Skills

Build reusable slash commands for K8s review, Docker audit, and team workflows.

/skills $ARGUMENTS /simplify
07

Infrastructure — k3s on DO

Provision a DigitalOcean droplet, install k3s, and deploy the app to Kubernetes.

k3s DigitalOcean
08

Hooks — Workflow Automation

Auto-format, notify, block dangerous commands — hooks fire at every lifecycle event.

hooks settings.json
09

MCP Servers

Connect Claude to GitHub, filesystems, and external APIs via the Model Context Protocol.

MCP mcp.json
10

GitHub Actions & CI/CD

Automate PR reviews, issue implementation, and code quality with Claude in CI.

@claude GitHub Actions
11

Sub-agents

Create specialized AI workers for security review, K8s validation, and parallel tasks.

/agents worktree
12

GitOps Finale — ArgoCD

The grand finale. Push to git, ArgoCD syncs, dark theme goes live on the internet.

ArgoCD /schedule MILESTONE
13

Advanced Patterns

Agent teams, plugins, headless mode, Desktop app — and where to go next.

teams /batch plugins

What You Need

Minimal prerequisites, minimal cost. That's the point.

Prerequisites

  • You know git and can open a terminal
  • Basic familiarity with Docker concepts
  • A GitHub account
  • A DigitalOcean account (new accounts get $200 free credit)
  • Curiosity and about 10 hours of focused time

No Kubernetes experience needed — we teach that as we go.

Cost Breakdown

Claude Pro subscription $20/mo
DigitalOcean droplet (s-2vcpu-4gb) ~$24/mo
Domain name (optional) ~$10/yr
Total for the course ~$25-45

Tear down the droplet after the course to stop charges. DigitalOcean new accounts get $200 free credit — enough to run this course for free.

Need help applying this to your real stack?

Get Mentoring →

Facts, pricing, and Claude Code features last verified on 2026-04-06. See official docs for the latest.