Deploy your inference pipelines to Kubernetes

Easily deploy inference steps in your pipeline by providing a link to your model. Kai integrates with all the popular model formats and storage providers.

Sometimes you need to perform pre or post processing steps within your pipeline. Kai makes it easy to deploy and integrate general services into your pipeline as well.

Kai simplifies your pipelines by separating the deployment of steps within your pipeline and the orchestration of those steps allowing you to scale workloads out indepdently in a Kubernetes native way.

By leveraging Kai you have full access to the Kubernetes ecosystem when it comes to sharing and distributing your inference pipelines. Tools like Helm, and Kustomize make it easy to package and version your entire pipeline.