Deploy scalable model serving on Kubernetes with Ray Serve.
npx skills add https://github.com/tylertitsworth/skills --skill ray-serve
Please help me install this Agent Skill. Skill: ray-serve Source: https://github.com/tylertitsworth/skills/tree/main/ray-serve Command: npx skills add https://github.com/tylertitsworth/skills --skill ray-serve
Ray Serve on Kubernetes addresses the challenge of scalable, production-grade model serving by orchestrating deployments, proxies, and autoscaling for ML workloads.
Apply a RayService CRD to deploy and scale your model-serving deployment on Kubernetes.
High-intent search queries and answers about installing and using this skill.
Scalable model serving on Kubernetes is deployed by applying a RayService CRD, which orchestrates Ray Serve deployments, routing, and autoscaling for production ML workloads.
Ray Serve manages production ML workloads by handling deployments, configurable proxies, batched requests, health checks, and multi-model pipelines across Kubernetes clusters.
Yes, autoscaling and batching for model inference are configurable through Serve config embedding, autoscaling parameters, and per-deployment options defined in the RayService CRD.
Yes, Ray Serve supports multi-model inference pipelines, streaming, and controlled routing to serve large language models and complex ML workflows on Kubernetes.
The best way to manage health checks and routing is using Ray Serve on Kubernetes, which embeds deployment defaults and per-deployment options directly within the RayService CRD configuration.
Yes, a Kubernetes cluster is required to apply the RayService CRD, which handles deployment defaults, autoscaling parameters, and health checks for scalable model serving.
truongnat...
tomzDeploy scalable LLM inference endpoints on Kubernetes with vLLM, TGI, or Triton.
ricableOrchestrate distributed model training across Ray clusters.
veerababumanyamDefine Traefik routing, KEDA autoscaling, and Prometheus observability for Kubernetes deployments.
