Johnder Perez

Free · Plain English · Self-paced

Computer Science for Vibe Coders

You can already build apps with Lovable, Cursor, Claude and friends. This course teaches you what's happening underneath, so you can keep the speed and lose the bugs, security holes and "I have no idea why this broke" moments.

Start: How Computers Actually Work → Get the CS Coach prompt

The idea: from prompting to engineering

Vibe coding (describing what you want and letting AI write the code) is a genuinely great way to prototype. But a prototype and a product are different things. Real apps have to stay secure, stay fast as data grows, and keep working when you change them next month.

That's what computer science and software engineering are for. You don't need a degree, and you don't need to stop using AI. You need a working mental model of the machine, the vocabulary to judge the code you're given, and a few habits that keep projects healthy.

This course is built around one principle: use AI as a tutor, not a crutch. Every track explains concepts in plain English, shows the specific mistakes AI tools tend to make, and gives you a hands-on checklist to prove you've actually got it.

The four tracks

Track 1 · 6 steps

How Computers Actually Work

How computers actually run your code

Track 2 · 6 steps

Thinking in Algorithms & Data

Why code is slow, and how to organize data

Track 3 · 6 steps

Writing Code That Lasts

Code that humans (and future you) can maintain

Track 4 · 7 steps

Building for Scale

Backends, databases and architecture

Plus a practical bonus track for people shipping real apps today:

Bonus track

Shipping real apps

Git branches, pull requests, CI, secrets, auth, migrations, database security, testing and deploys.

 Habits

Rules of engagement

Four anti-vibe-coding rules that build real skill.

 Tool

The CS Coach prompt

Turn any AI assistant into a patient CS tutor. One-click copy.

How to use this course

  1. Set up your tutor. Copy the CS Coach prompt into Claude, ChatGPT, Grok or your coding tool's custom instructions.
  2. Go track by track. Each track is split into short steps with a quick quiz, a hands-on challenge and a milestone checklist. Ticks are saved in your browser.
  3. Apply it to your own app. Every track lists mistakes to look for. Hunt for them in something you've built.
  4. Ship carefully. Use the Shipping real apps checklist before you put anything with real users or real data online.

Who this is for

Founders, designers, marketers, realtors, students, anyone who has built something with AI and wants to understand it well enough to trust it. If you've ever pasted an error back into a chat five times in a row, this is for you.

Estimated pace: a few hours a week. The first track takes about 4–6 weeks alongside CS50x; the whole path takes a few months. Go at your own speed.