Mesh LLM enables distributed AI inference by pooling GPUs across machines via iroh

How Mesh LLM pools existing GPU resources across machines into a single OpenAI-compatible API, built on iroh.