Kubeflow

Kubeflow is an open-source AI platform that helps teams build, manage, and scale data, AI, and machine learning workloads on Kubernetes.

They provide portable ML workflows, interactive notebooks, distributed AI training, and scalable infrastructure for teams working across different environments.

Their standout capability is Kubeflow Trainer, which supports scalable LLM fine-tuning and model training across frameworks such as PyTorch, HuggingFace, JAX, and XGBoost.

They also offer automated machine learning, hyperparameter tuning, model management, pipelines, and a centralized dashboard through modular open-source projects.

Kubeflow is fully open source, allowing teams to use, modify, and self-host its components without a paid plan or free trial requirement.