deployKF Enterprise will 🚀 soon
Express your interest, share feedback, and unlock a permanent discount. (subject to terms)
Start for free with the open-source project and deploy Kubeflow and KServe.
Scale groups of Airflow workers to meet demand. Only consume GPU and server resources when required.
Run fewer workers without disrupting critical or long-running jobs. Define custom rules for your specific workload requirements.
Reduce network congestion and improve security with isolated KServe model deployments.
Create isolated network paths (Istio Gateways) for model serving, or deploy to a separate Kubernetes cluster for even more isolation.
Real-time collaboration (similar to Google Docs) for JupyterLab.
Allow multiple people to work on the same notebook environment without interrupting each other.
Off-cluster SSH access into Notebook Pods.
Use remote execution features to run native IDEs like PyCharm and VS Code from your laptop.
Create new notebook images without Docker.
Make changes to a running notebook (add code or install packages) then use the snapshot button to make a new image.
The Admin Hub helps configure, integrate, and troubleshoot your deployKF platform. It is a self-hosted web interface that runs alongside deployKF.
For example, to integrate deployKF with a database, it can walk you through the required configs and test that you have set them correctly.
Automatically define users, groups, and profiles by syncing them from an external identity provider.
For example, rather than manually defining all users, groups, and profiles in the deployKF values, they can be automatically pulled from systems like Active Directory (LDAP) and Okta.
Reference Patterns
Documentation and reference implementations for common AI/ML use-cases.
For example, how to use MLflow (Model Registry) to track training runs and manage model deployments using GitOps.
Use a single deployKF dashboard across multiple Kubernetes clusters.
For example, a dashboard on one cluster can provide access to Airflow instances that run on a separate cluster from the dashboard.
Easily add your branding to deployKF, with custom logos and colors.
Offline Deployment
Run deployKF in offline/air-gapped environments with fewer steps and complexity.
An official Pulumi provider to manage deployKF.
An official Terraform provider to manage deployKF.
deployKF Enterprise will be available worldwide, starting in the United States and other western countries.
Please join the waitlist to express your interest and unlock a permanent discount. (subject to terms)
deployKF Enterprise runs on your Kubernetes cluster, so it's not a managed service in the traditional sense.
This approach gives you full control and flexibility to run anywhere in the cloud or locally.
We aren't quite ready to publish pricing information, please contact us after joining the waitlist for more information.
Migrating from deployKF to deployKF Enterprise is very easy. In most cases, you can migrate in-place with minimal downtime.