Who this series is for

You already write software. You’re comfortable with a code editor, a terminal, and version control. You’ve probably used GitHub Copilot’s autocomplete or asked ChatGPT to explain an error message. This series picks up from there.

If you can clone a repo, install dependencies, and open a pull request — you have everything you need to follow along. If terms like “PR,” “CI,” or “environment variable” are unfamiliar, you’ll want to learn the basics first.

This series is not about making an assistant produce the most code. It is about using AI without lowering the engineering bar. The workflow assumes that you still own the design, the trade-offs, the security posture, and the final diff. The assistant can draft, search, test, explain, and automate. You remain accountable for what ships.

The posture I recommend is simple: treat every AI output as a hypothesis. Verify it against the spec, the codebase, the tests, and the risk model before you accept it.

Required tools

Install these before starting Chapter 1. Every hands-on exercise in the series assumes they’re available.

1. Visual Studio Code

VS Code is the primary editor throughout the series. As of version 1.99 (April 2025), GitHub Copilot is built in — no separate extension required.

• Download: code.visualstudio.com
• Minimum version: 1.99+
• Verify: Open VS Code → Help → About → check the version number

2. Node.js

Most code examples use JavaScript or TypeScript. Node.js is also needed to run MCP servers and various CLI tools.

• Download: nodejs.org (use the LTS version)
• Minimum version: 22+
• Verify:

node --version
# Expected: v22.x.x or higher

3. Git

You need Git for version control and to interact with GitHub repositories.

• Download: git-scm.com (or install via your OS package manager)
• Minimum version: 2.39+
• Verify:

git --version
# Expected: git version 2.39.x or higher

4. A package manager: npm or pnpm

npm ships with Node.js. Several chapters also use pnpm in examples — either works, but if you want to follow the examples exactly:

# npm is already installed with Node.js
npm --version

# To install pnpm (optional, but used in many examples)
corepack enable
corepack prepare pnpm@latest --activate
pnpm --version

5. GitHub account + Copilot access

You need a GitHub account with access to GitHub Copilot.

• Create an account: github.com/signup (if you don’t have one)
• Copilot access: The Free tier works for most exercises. Copilot Pro is recommended — it removes the monthly completion limits that might interrupt your workflow during longer exercises
• Verify: Open VS Code, click the Copilot icon in the status bar, and sign in with your GitHub account

Recommended (not required)

These tools appear in specific chapters. You don’t need them upfront — install them when the chapter calls for them.

Each chapter that introduces an optional tool includes installation instructions in context.

Knowledge this series assumes

The series teaches AI-assisted engineering workflows — not programming fundamentals. Here’s what you should be comfortable with before starting:

Must-have

• Terminal basics — navigating directories (cd), running commands, reading output. You don’t need to be a shell expert, but you shouldn’t be afraid of the command line.
• JavaScript or TypeScript — most code examples use TypeScript. You need to read and understand basic syntax: functions, types, async/await, imports. You don’t need to be an expert — the AI will help.
• Git fundamentals — clone, commit, push, branch, merge. You should know what a pull request is and how to create one.
• GitHub navigation — creating repositories, opening issues, reviewing PRs on github.com.

Helpful but not required

• REST API concepts — HTTP methods, status codes, JSON request/response. Chapters 2, 10, and 18 use API examples extensively.
• Testing concepts — what unit tests, integration tests, and E2E tests are. Chapter 10 explains the workflow but assumes you know the basic taxonomy.
• CI/CD basics — what GitHub Actions does and why you’d want automated pipelines. Chapters 14 and 15 touch on automation and CI-adjacent concepts.
• Another programming language — some examples use Python, Go, or SQL to show that the workflows are language-agnostic. You don’t need to know these languages — the principles transfer.

Mindset required

You will get more from the series if you are willing to slow down at the right moments. AI-assisted work fails when the developer optimizes for generation speed instead of delivery quality. The habits that matter most are old engineering habits: write down intent, keep changes small, run the checks, review the diff, and measure what happened.

That matters because the evidence is mixed. Controlled experiments show real speedups on bounded tasks, while studies with experienced developers on mature codebases also show slowdowns when review and correction cost dominate. The workflow in this series is designed for that reality.

Operating system

The series is OS-agnostic. Keyboard shortcuts are given for both macOS (Cmd) and Windows/Linux (Ctrl). Installation commands cover Homebrew (macOS), apt (Linux), and WinGet (Windows) where applicable. Hook scripts show both Bash and PowerShell variants.

If you’re on Windows, you’ll have the smoothest experience using WSL 2 (Windows Subsystem for Linux) for terminal commands, though it’s not strictly required.

Quick setup checklist

Run these commands to verify everything is ready:

# Check VS Code (run from terminal)
code --version
# Expected: 1.99.x or higher

# Check Node.js
node --version
# Expected: v22.x.x or higher

# Check Git
git --version
# Expected: git version 2.39.x or higher

# Check npm
npm --version
# Expected: 10.x.x or higher

# Check GitHub authentication (optional but helpful)
gh auth status
# Expected: Logged in to github.com

If all commands return valid versions, you’re ready. Open Chapter 1 and let’s get started.

Next up: Ch 1 — Introduction & Overview — the tool landscape, what AI can and can’t do, and the workflow that ties this series together.