Appearance
Day 1: The Vibe Coding Workflow
You already know what vibe coding looks like from the scenarios in the first two lessons. Someone describes a feature in plain English, and a working system appears. The question is not whether it works. The question is how to make it work consistently, across days, across sessions, across projects where the stakes are higher than a Saturday afternoon side hack.
Today is about the mental models that separate a lucky first prompt from a repeatable workflow. Project managers carry around a head full of concepts that map directly onto vibe coding: a spec is a system prompt, a task breakdown is a sequence of prompts, dependencies are the order you describe things. The LLM underneath you has preferences too, code patterns it produces more reliably when you structure your instructions a certain way. And the prompts themselves need engineering when you are not throwing away the output after one session but building something that needs to stay coherent across a week of work.
By the end of today, you will have the framework: design thinking that keeps a project coherent, code patterns that produce consistent output, and prompting strategies that scale beyond a single afternoon.
What you will cover
- Vibe Coding in Practice -- real scenarios, real outcomes. The five patterns every successful vibe-coded project follows, from describing outcomes to iterating fast.
- Your Vibe Coding Toolkit -- Claude Code setup, CLAUDE.md for persistent project memory, custom slash commands, file injection, auto-approve mode, and git as your undo button.
- Systems Design & Project Management for Vibe Coding -- how traditional PM concepts like requirements, scope, task breakdown, and dependencies map directly onto vibe coding. A spec becomes a system prompt. Architecture becomes the context you feed the agent.
- How LLMs Like to Code -- the code patterns LLMs produce more reliably: pure functions, descriptive names, explicit error handling, small files. Concrete before/after examples in Python and JavaScript.
- Prompting Patterns That Scale -- system prompts as project specs, context pinning, caching strategies, and structuring prompts so an agent stays aligned across multiple sessions and days of work.
When these five lessons click, you are not a person typing clever prompts. You are a project leader directing an AI development process. The distinction is real, and it is what Day 2 builds on.