How to use ai-engineering-hub: clone, pick a project and learn

What ai-engineering-hub is and who it suits

The repository is a large set of tutorial projects that show how to build with LLMs, RAG and AI agents. The README describes it as a resource for learning and building with AI, with projects across all skill levels.

It works best if you want working examples to study and adapt. It is a poor fit if you want a single ready-made application to install and use, or if you are not comfortable with Python and notebooks, since most projects assume both.

• Good fit: developers who want RAG or agent examples to adapt
• Good fit: beginners who want guided projects
• Good fit: students seeing how LLM tools fit together
• Poor fit: people who need one finished app
• Poor fit: non-programmers

Install prerequisites: Git and Python

You need Git to copy the repository to your machine. Most projects also need Python, but the README excerpt does not state which Python version to use, so check the individual project folder for its requirements.

The README does not say how each project is maintained, so expect that some examples may need small fixes to run with current libraries.

• Git installed on your computer
• Python (version not stated in the README excerpt)
• Familiarity with Jupyter notebooks is helpful

Clone the repository

Open a terminal and run the clone command from the README. This creates a local folder containing every project.

Choose a project by difficulty

The README groups projects into beginner, intermediate and advanced lists. The beginner list focuses on single components and straightforward implementations, such as OCR apps and simple RAG workflows. Intermediate projects cover multi-component systems and agentic workflows, and advanced projects cover complex systems and fine-tuning.

If you are new, the README suggests starting with the AI Engineering Roadmap, then the beginner projects, then intermediate and advanced ones.

• Beginner: for example Simple RAG Workflow, which uses LlamaIndex and Ollama
• Beginner: for example Document Chat RAG, which uses Llama 3.3
• Intermediate: for example Agentic RAG, which adds web fallback to document search
• Intermediate: for example AutoGen Stock Analyst, which uses Microsoft AutoGen
• Advanced: fine-tuning and production-style projects

Run a project and adapt it

Each project lives in its own folder with its own notebook or app. Open the folder, read its README or notebook instructions, and follow the setup steps it gives, since those steps are specific to that project and are not in the excerpt.

To adapt a project, change one component at a time, such as the model, the embedding step or the document source, and check the output after each change. This makes it easier to see how each piece of the stack affects results.

• Read the project-level instructions before installing anything
• Change one component at a time
• Compare results across models using the comparison projects

Compare with other frameworks

The repository is useful for seeing examples side by side, but it is not a framework itself. If you want a library to build on, the comparable projects it is often weighed against are LangChain, LlamaIndex, Microsoft AutoGen and CrewAI. Several examples in the hub use LlamaIndex, AutoGen or CrewAI directly.