Members-Only
Recent Talks & Demos are for members only
You must be an AI Tinkerers active member to view these talks and demos.
Krayne: Self-Serve Ray Clusters
Learn how Krayne simplifies Ray cluster setup for researchers, bypassing Kubernetes YAML. See how simple commands translate into complex KubeRay resources for seamless remote Ray execution.
Krayne, a lightweight CLI, Python SDK, and interactive TUI (ikrayne) that provisions self-serve Ray clusters for AI researchers who dislike wrestling with Kubernetes manifests.
For the demo, I will show how to bypass K8s YAML entirely to instantly bootstrap a Ray cluster using a simple command (krayne create) or through our interactive terminal interface (ikrayne).
Then, I’ll peek under the hood to show the actual technical implementation: how the tool translates these simple inputs into the complex KubeRay Custom Resources and API calls required to orchestrate the pods behind the scenes.
Krayne abstracts KubeRay via Pydantic-validated CLI, TUI, and Python SDKs.
Python SDK/CLI wrapping KubeRay for programmatic Kubernetes Ray cluster orchestration.