Inference Starts Here: Open Inference. Real Impact. One day.
Join The Open Accelerator for a fast-paced, single-day hackathon focused on using an open-source AI inference stack! Come to see how models are served!
This isn't a tutorial—this is your chance to build on vLLM alongside core committers from Red Hat and IBM who will be right there as your mentors. Tackle real open-source issues and build production-viable applications across our five challenge tracks that focus on optimizing inference cost, creating a RAG pipeline, using speculative decoding, maximizing performance and bring your own problem.
We use a "one theme, three skill lanes" approach (Starter, Builder, and Deep Tech), so whether you're shipping your first model, vibe-coding a frontend with Cursor, or optimizing token throughput at the kernel level, there is a path for you. Vibe coding isn't just allowed—it's highly encouraged!
Registration includes access to shared GPU compute, pre-quantized LLM models, food, and a great collaborators.
Registration acceptances will be sent out the week of April 20th along with some pre-reads on the tooling, proposed challenges to conquer, and the precise location.
Teams of 3-5 will be formed onsite by those who show up to build!
Come for the tech, stay for the community. Let's see what you and your machine can do!