Google launched AI Hub for its Cloud platform which is available for alpha and Kubeflow Pipelines. This product was launched to help developers in managing and productionizing large scale machine learning modules with ease. Company opened its Kubeflow Pipelines as open-source project and it will enable business to use AI Hub platform for managing machine learning workflows.
According to the Google Cloud senior director of product management Rajen Sheth there are only few thousands of Machine Learning Engineers in the world who can bring concept to production, but there are millions of data scientists and millions of developers. Google AI Hub Kubeflow Pipelines is designed to fill this gap. These tools are aimed to empower data scientist and machine engineers in managing complex machine learning pipelines with ease.
According to Sheth many Company’s machine learning investments are getting wasted due to scarcity of Data Engineers with right skillsets which can deploy the ML Models in production.
“One of the biggest problems we’re seeing right now is companies are now trying to build up teams of data scientists, but it’s such a scarce resource that unless that’s utilized well, it starts to get wasted,” Sheth said.
Sheth futher added “One observation we’ve seen is that in probably over 60 percent of cases, models are never deployed to production right now. So we’re building a number of things to hopefully help cure that. ”
Google come up with the Kubeflow Pipelines, which is a composable layer that provides API to deploy and run machine learning modules over production clusters. It can be used to easily design and stich ML components to make fully functional machine learning work flow, which can be executed on the distributed cluster.
The Kubeflow is an open source project from Google which as released as open source project this year on the Github. This project can be used with the Kubernetes containers for designing and deploying machine learning workflow on distributed cluster. Kubeflow allows business to make flexible system for training AI Modules with on-premise data or on the cloud platform. With Kubeflow developers can use the pre-existing Tensorflow libraries, add own library and make their ML workflow.
Kubeflow uses the libraries from TensorFlow Extended (TFX) which developers can use in their program. The TensorFlow Extended (TFX) is set of libraries was used at Google to build machine learning components. So, developer has access of most advanced libraries for developing, deploying and machine learning modules for their businesses.
