"Why did the AI write this code?" or "Why can't it remember what I just said?" The frustration you encounter when using AI tools like Claude Code stems from not knowing how the tool actually works. This course flings open that black box. Over 25 lessons, we will build our own Coding Agent from scratch that functions exactly like real-world professional tools—giving the LLM 'eyes' to see files (Read), 'hands' to fix code (Edit), and 'feet' to execute commands (Bash). This isn't just about learning how to use a library. You will gain hands-on experience by implementing code to understand how AI perceives project structures, why it loses its memory during long conversations, and how to solve those issues technically (Sliding Window, Summary). Through this process, you will evolve from being a mere 'user' of AI tools to an 'AI provider' who understands the underlying principles and can customize and evaluate tools to fit your specific needs.

This course is a practical program where you will master Kotlin-based AI agent development by building your own coding agents. Using Koog, an AI framework built with Kotlin, you will learn step-by-step how to implement coding agents similar to Claude Code, Cursor, and Junie. Koog is more than just a simple server-side framework; it integrates seamlessly with Spring and offers the scalability to connect with various environments—including Android, iOS, Desktop, and Web—through Kotlin Multiplatform (KMP). By building intelligent agents that can read, analyze, modify, and generate code, you will gain hands-on experience in AI agent architecture, strategy, context handling, and tool integration.