welcome
Thank you for purchasing the MEAP of Self-Hosting LLMs at Scale.
This book is for engineers who have been asked to run large language models on infrastructure their company controls. To get the most out of it, you should be comfortable with Python scripts, the Linux command line, and basic Docker. It helps to know what Kubernetes is, but you don't need production Kubernetes experience. You also don't need any background in GPUs, inference engines, or machine learning theory. The book treats models as artifacts to be served, and it explains the hardware and the serving engine from the ground up.
I work on the infrastructure behind the Hugging Face Hub, and I have seen many teams try to move from a hosted model API to their own deployment. The questions they ask are very similar. How much GPU memory does this model need? Which model should we choose, and can we legally use it? Why is the endpoint slow, and how many users can one GPU serve? What does it cost compared with the API? This book collects the answers in one place.