Are the training and inference running on Kubernetes?

[pacoxu]

Most mentioned tech are about Model, CUDA and deeper layers.

Are the training and inference running on Kubernetes? Does deepseek using some native components like

1. https://github.com/kubernetes-sigs/lws?
2. https://github.com/vllm-project/vllm/ or https://github.com/sgl-project/sglang?
3. https://github.com/kubernetes-sigs/kueue?

Or some other native solutions(open source or not)?