DEVOPSdigest

D2iQ introduced KUDO for Kubeflow to simplify and accelerate machine learning (ML) deployments on Kubernetes.

An enterprise-ready distribution of open source Kubeflow, D2iQ KUDO for Kubeflow accelerates time-to-market for ML workflows by reducing the complexity of provisioning and managing all of the moving pieces. Bundled with other ML tools such as Spark and Horovod, KUDO for Kubeflow delivers an end-to-end secure, scalable and portable ML platform that empowers data scientists and ML engineers to more quickly and consistently build, deploy and run workflows in Day 2 operations.

KUDO for Kubeflow empowers organizations with a platform that provides standardized best practices and tools for running machine learning on Kubernetes. By removing the complexity of setting up ML development and production environments, KUDO for Kubeflow enables organizations to improve the productivity of data science teams at a much lower cost. Data scientists can leverage GPUs and MLOps to speed up the process of training, tuning and deploying models, regardless of the underlying infrastructure, reducing the costs and risks associated with manual setups. ML engineers can now deploy and train ML models at scale, all on a single platform.

"Taking ML workflows from development to production is filled with challenges, as discrepancies between the environments, monolithic architectures, and lack of portability and scalability are common when trying to deploy a model into production," said Chandler Hoisington, SVP Engineering and Product, D2iQ. "D2iQ KUDO for Kubeflow enables organizations to develop, deploy, and run entire ML workloads in production at scale, while satisfying security and compliance requirements. This enables data scientists and ML engineers to run their entire ML stack with much higher velocity on Kubernetes infrastructure."