Vibe Coding Course — Build Real Software with AI
Vibe coding has changed the barrier to software creation. But there is a major difference between using AI to write code and knowing how to build software with AI — the second requires understanding requirements, architecture, project context, debugging, testing, security and deployment. This track teaches the second, across eight sessions at a typical pace of eight weeks, starting from beginner-to-intermediate level.
- Format: 8 sessions · 8 weeks · Beginner to intermediate · prerequisites: helpful but not required.
- Module 1 — AI coding environment setup: How modern AI coding tools fit into a real software workflow — Cursor, Claude Code, GitHub Copilot, Lovable, GitHub, Vercel and Supabase, with the exact stack chosen per project. You build: A properly configured AI-assisted development environment.
- Module 2 — Prompting for software development: Software prompts need far more structure than ordinary prompts — including when not to ask an AI tool to rewrite an entire project, because large, poorly scoped instructions create large, poorly understood changes. You build: A reusable software-development prompt framework.
- Module 3 — Building a full-stack application: You build a real project that solves a real problem — an internal tool, an AI assistant, a SaaS MVP, a research tool or a customer-facing product. You build: A working full-stack application.
- Module 4 — Authentication, database, payments and email: The infrastructure behind real products — the pieces AI-generated applications most often gloss over. A prototype can look impressive while still being unusable. You build: A functional application with real user flows, not just a polished interface.
- Module 5 — Refactoring and working with existing code: Using AI against existing applications, legacy systems and large repositories — starting by asking AI to explain a system before modifying it, which is the step most developers skip. You build: A refactoring or improvement plan for an existing project.
- Module 6 — Testing, CI and AI-assisted code review: Generated code still needs testing. AI can generate a test, but the test must actually test the intended behaviour. You build: A testing and review workflow for your application.
- Module 7 — Performance, SEO, accessibility and production deployment: A functional application is not automatically a successful one. Deployment typically uses platforms such as Vercel and Supabase depending on architecture. You build: A production-ready deployed application.
- Module 8 — Launch, feedback and iteration: Getting the application in front of users, then improving it on evidence rather than instinct. You build: A launched product with a working iteration loop.
- Each 60-minute session: 10 minutes goal and review, 35 minutes live building, 10 minutes documenting the playbook, 5 minutes setting the next milestone.