Kubernetes for AI Inference: Serving Models at Production Scale
Kubernetes has quietly become the de facto control plane for production AI inference. This post walks through the architecture, the GPU plumbing, and the patterns teams use to serve models that actually meet SLOs.