The Kaptain AI/ML Add-on, our enterprise MLOps solution, leverages open-source KubeFlow plus everything else you need to get your models from notebooks into production faster and more robustly. The Jupyter notebooks that data scientists use are pre-installed with many top libraries such as SciPy and Keras, training frameworks PyTorch, Tensorflow, and Scikit-learn, and AutoML with Katib.

Kaptain provides many enhancements, integrations, and user experience improvements on top of open source KubeFlow. Kaptain is easily deployed via helm, an improvement on top of Kubeflow. The Kaptain SDK, a Python API that abstracts away underlying Kubernetes details, allows data scientists to focus on building, tuning, and deploying models. Kaptain also implements MLFlow for experiment metadata tracking.

The Kaptain AI/ML add-on prioritizes GPU workloads, with out-of-the-box support for Nvidia GPUs, Triton serving, and tight integration with Nvidia's GPU catalog (NGC) for easy access to pre-trained models. Kaptain is certified on Nvidia's DGX platform.

Kaptain provides multiple security advantages - includes OIDC provider support, CVE mitigation, and robust air-gapped testing.