As large language models (LLMs) become increasingly integral to modern applications, organizations are exploring self-hosted solutions to gain greater control over performance, data privacy, and customization. This session offers a quick introduction to LLMs, focusing on the motivations and methodologies for deploying them in self-hosted environments using open-source tools. Attendees will gain practical insights into the architectural considerations, hardware requirements, and performance trade-offs involved in building and maintaining a local LLM stack. The session will conclude with a live demonstration showcasing the deployment of an open-source LLM, highlighting key steps and best practices for getting started.