Hands-On with vLLM: Fast Inference & Model Serving Made Simple

Tired of slow inference and complex serving pipelines? Join us for a **live hands-on demo of vLLM**, the high-performance inference engine designed for large language models. ​In this session, you’ll learn: * ​How to install and configure **vLLM** step by step * ​Best practices for serving models efficiently with **dynamic batching and PagedAttention** * ​How vLLM compares to traditional serving frameworks like TGI and Hugging Face Inference * ​Tips for running vLLM locally and scaling on the cloud ​This is a **practical, no-fluff workshop**—you’ll walk away with a running model served via vLLM and the know-how to deploy your own in production. ​🔹 **Format:** Live coding + Q&A 🔹 **Who’s it for:** AI engineers, MLEs, founders, and anyone curious about deploying LLMs at scale 🔹 **Takeaway:** A working vLLM setup and a deeper understanding of efficient LLM serving