在 Kubernetes 上构建经校验的 NVIDIA Dynamo 运行时
原标题：Build a Validated Kubernetes Runtime for NVIDIA Dynamo
AI 摘要
NVIDIA 发布技术博客，介绍如何通过 NVIDIA AI Cluster Runtime（AICR）在 Kubernetes 上部署并验证 NVIDIA Dynamo 分布式推理框架。AICR 将 Dynamo 及其依赖的 GPU Operator、DRA 驱动、KAI Scheduler、Grove、Prometheus 等基础设施打包为带版本锁定和校验的 recipe，使平台团队可重复部署。Dynamo 1.4 起前端与 worker 通过 Kubernetes 资源发现，默认不再使用 NATS，并支持基于 KV 缓存感知的路由与分离式 serving。
正文节选
Teams building AI applications need a repeatable path from a model configuration to a working inference endpoint. NVIDIA Dynamo is a distributed inference framework that coordinates model serving across GPU workers. Deploying it on Kubernetes is only a first step. A successful model deployment requires more: GPU drivers, resource allocation, scheduling, networking, and monitoring must all work together. Resolving these dependencies takes time away from testing model behavior, serving performance