“NVIDIA contributes to open source projects including Dynamo for disaggregated inference, as well as runtimes like VLLM and SGLang.”