Build LLM, RAG, and AI agent systems with 93+ practical code projects
This repository offers a place to learn complex AI concepts by running actual code rather than just studying theory. It collects step-by-step practical examples for building Large Language Models (LLM), Retrieval-Augmented Generation (RAG), and AI agents. From writing your first line of code as a beginner to deploying systems in production, you can find all the necessary pieces here.
Learning paths by difficulty
The repository organizes projects into three stages tailored to the learner's level. The beginner stage starts with single-component tasks like OCR apps or simple RAG implementations to establish the basics of AI engineering. The intermediate stage covers agent workflows and advanced RAG systems that combine multiple components, teaching how to solve complex problems encountered in practice. The advanced stage delves into model fine-tuning, production deployment, and cutting-edge infrastructure to deepen expertise. This structure allows learners to select projects matching their current skill level and build competence gradually.
Project structure by core technical area
Within each difficulty level, projects are categorized by specific technical areas, allowing you to focus deeply on desired topics. The collection includes various fields such as OCR and vision, chat interfaces, and multimodal media processing, with Model Context Protocol (MCP) projects widely distributed across the intermediate and advanced levels. It also features topics reflecting current AI trends, such as model comparison and evaluation, voice and audio processing, and advanced agent systems. Each project is provided in Jupyter Notebook format, making it easy to run code and verify results.
Tools and frameworks for practical application
The projects in this repository provide examples that go beyond simple demos and can be applied to real-world projects. It includes rich implementation examples using tools and frameworks widely used in current AI engineering, such as LlamaIndex, Ollama, CrewAI, and AutoGen. This allows readers to understand not just how to use specific libraries or frameworks, but also the structure of how they combine to solve real problems. Additionally, the mix of projects runnable in local environments and examples of cloud-based service integration helps improve understanding of various deployment environments.
Checklist before starting to learn
This repository is released under the MIT license, allowing free use for learning and commercial purposes. However, some projects may require specific hardware specifications or GPUs, so you should verify that your development environment meets the requirements before attempting advanced projects. Also, since the models and library versions used can differ from project to project, it is advisable to carefully review each project's README or code comments before execution. As this repository serves as dynamic learning material to keep up with the rapidly changing AI field, it is effective to check for the latest updates regularly while learning.
