Kubeflow SDK v0.4.0: Model Registry, SparkConnect, and Enhanced Developer Experience

Unified Model Management: The Model Registry Client Usage Example Distributed AI Data at Scale: SparkClient & SparkConnect Usage Example A New Home for Documentation Infrastructure & Breaking Changes Better Isolation with Namespaced TrainingRuntimes Furthering Parity Between Local and Remote Execution Required: Upgrading to Python 3.10+ What’s Next for Kubeflow SDK Get Involved! Unified Model Management: The Model Registry Client Usage Example Usage Example Distributed AI Data at Scale: SparkClient & SparkConnect Usage Example Usage Example A New Home for Documentation Infrastructure & Breaking Changes Better Isolation with Namespaced TrainingRuntimes Furthering Parity Between Local and Remote Execution Required: Upgrading to Python 3.10+ Better Isolation with Namespaced TrainingRuntimes Furthering Parity Between Local and Remote Execution Required: Upgrading to Python 3.10+ What’s Next for Kubeflow SDK Get Involved! Explore the full documentation at sdk. kubeflow. org With KubeCon just around the corner, we are pleased to announce the release of Kubeflow SDK v0.4.0. This release continues the work toward providing a unified, Pythonic interface for all AI workloads on Kubernetes. The v0.4.0 release focuses on bridging the gap between data engineering, model management, and production-ready ML pipelines. The Kubeflow SDK now covers most of the MLOps lifecycle – from data processing and hyperparameter optimization to model training and registration: Highlights in Kubeflow SDK v0.4.0 include: Model Registry Client for managing model artifacts, versions, and metadata directly from the SDK. SparkClient API with SparkConnect support for interactive data processing Namespaced TrainingRuntimes for improved isolation and multi-tenant platform management Dataset and Model Initializers enabling better parity between local and Kubernetes execution A new Kubeflow SDK documentation website with examples, and API reference Minimum Python version updated to Python 3.10 for improved security, typing, and runtime performance Managing model artifacts, versions, and metadata across experiments has historically required stitching together multiple tools outside of your training code. In v0.4.0, the SDK introduces ModelRegistryClient – a Pythonic interface to the Kubeflow Model Registry, available under the new kubeflow. hub submodule. ModelRegistryClient kubeflow. hub The client exposes a minimal, curated API: register models, retrieve them by name and version, update their metadata, and iterate over what’s in your registry – all without leaving the SDK. It integrates directly with the Model Registry server and supports token auth and custom CA configuration for production clusters.