Introduction to Vibe coding for non-programmers

The course is intended for anyone who wants to create their own applications with the help of AI - for beginners with no programming experience, managers, salespeople, specialists in operations, marketing, finance or HR, analysts, product owners, testers and developers who want to systematically involve AI coding agents in their work.

Practical exercises take place in Claude Code, but the principles learned can also be used in other agent development tools.

• understand the basic building blocks of a program
• create a simple algorithm using AI
• verify the correctness of the code using tests and unit tests
• use Git as a secure project history
• create and gradually expand the application using an AI coding agent
• continue working on an existing project across sessions
• systematically find and fix errors
• work with the limits of the AI ??agent's context
• manage the work of the AI ??using project instructions and structured documentation
• plan major changes before their implementation
• prepare the project so that it can be followed by a new AI agent, a new session or another developer

• Basic computer work
• Experience with programming or AI tools is not necessary
• Own account with access to Claude Code (a paid Claude Pro, Max, Team or Enterprise subscription, or a Claude Console account with API credit)

Basic building blocks of the program

A simple algorithm created using AI

• Entering an AI chat task
• Transferring code to the program and running it
• Editing the result

Testing and unit tests

• Why test code created by AI
• Manual test and unit test
• Running tests

Git as a safe project history

• Diff and commit
• Change history
• Reverting to the previous version

The first simple application using the AI ??coding agent

• Running Claude Code in a project
• Entering and creating an application
• Checking and approving changes
• General principle of the agent's work directly with project files

Working across sessions and continuing in an existing project

• Named sessions in Claude Code
• Returning to work in progress
• What the agent must find out again when continuing the project

Debugging

• Reproducing the error
• Finding the cause
• Repair
• Result verification

Multi-file applications with persistent data

• Splitting the application into multiple files
• Saving and loading data

Gradual expansion of the application using an AI coding agent

• Small changes instead of large interventions
• Continuous verification
• Commit after each functional change

Limits of context and work across sessions

• What is the AI ??agent context and why is it limited
• What can be lost during long work or a new session
• How to ensure that the agent knows important information about the project

Project instructions for an AI agent

• Why use persistent project instructions
• CLAUDE.md as their practical implementation in Claude Code
• Creating and maintaining project instructions
• Transferring the same principle to other AI coding agents

Project rules and how to work with a project

• Architectural rules
• Project conventions
• Commands for launching the application
• Commands and rules for testing

Structured project documentation

• Splitting information into requirements.md, architecture.md and testing.md
• Separating requirements, technical design and testing rules
• Referencing documentation from project instructions

Planning before implementing major changes

• Drafting a plan before changing the code
• Checking the proposed procedure
• Implementing according to the agreed plan

Handing over and continuing the project

• README.md and how to start the project
• Known limitations
• Handing over the project to a new session, another AI agent or another developer

Related courses

Programming and Algorithmization Basics [PRG]