Makes it Easy for Everyone to Develop, Deploy and Manage a Portable, Distributed and Scalable ML system on Kubernetes. It reduces the development cycle timeline for AI/ML models.
Streamlined Machine Learning Deployments
Deploy, scale and manage your machine learning services with Kubeflow and Kubernetes.
Kubeflow is an open-source project that aims to make the process of deploying machine learning workflows on Kubernetes simpler, more portable, and more scalable. It includes support for TensorFlow Serving Containers for exporting TensorFlow models to Kubernetes, TensorFlow model training, and services for creating Jupyter notebooks. Kubeflow pipelines allow for efficient deployment and management of end-to-end machine learning workflows. End-to-end streaming using Kubeflow, Kafka and Redis can help create complex models at scale, quickly.
Bring automation to your processes - discovering the right parameters can take several days, Kubeflow removes those manual steps for you.
Improve the feedback loop enables it's users to conduct much more research while improving the safety and quality of your models
The Kubeflow pipelines SDK allows you to run a pipeline directly, this allows for rapid experimentation
From Notebook to Kubeflow Pipelines
For teams that deal with machine learning (ML), there comes a point in time where training a model on a single machine becomes untenable. This is often followed by the sudden realization that there is more to machine learning than simply model training.
There are a myriad of activities that have to happen before, during and after model training. This is especially true for teams that want to productionize their ML models.