llm_engineering

This repository holds the code and notebooks for an eight-week course called "LLM Engineering" by Edward Donner, available on Udemy. The course teaches how to build real applications using large language models (AI systems like GPT and Claude), aimed at people who want to go from beginner to confident practitioner through hands-on projects.

The course is structured as eight weekly modules, each building on the last. It starts with running a local AI model on your own computer using a free tool called Ollama, then progresses through using hosted AI APIs from providers like OpenAI and Anthropic, and eventually reaches Week 8 where students build an autonomous AI agent that can take actions on its own. From Week 3 onward, some exercises use Google Colab, a free cloud environment that provides access to GPU hardware for running larger models.

API costs are kept intentionally small. The instructor estimates a few dollars total across the entire course, and provides free alternatives (Ollama for local models, Google Colab for cloud GPUs) for every paid step. For the optional Week 7 section that involves training a model, the instructor personally spent about $10 but notes it is not required.

The repository was fully refreshed in December 2025 with new course content. Students who want to follow along with the original video recordings can check out an older branch called "original" via git. Each weekly folder contains Jupyter notebooks with code that students are expected to run, modify, and experiment with. The course places emphasis on doing rather than watching: students are encouraged to tweak every example and submit their own solutions as pull requests.

The repo is organized into week folders, a guides directory with supplementary notebooks, and a setup directory with installation instructions for all platforms.