UbiOps

0 Reviews

0 Followers

1 Like

5 Deploys

Run your data science code live in the cloud

Page last modified

April 11 2024

UbiOps is an easy-to-use deployment and serving layer for your data science code. Run your Python & R models and scripts live and use them from anywhere at any time. UbiOps is currently used as a backend to optimize heat networks by Gradyent. Furthermore BAM energy systems uses UbiOps to predict energy usage of buildings and builds their services on top of that.

Pros & Limitations

Up to 80% reduced time to market

No in-depth IT knowledge required

Support for complex pipelines

Automatic API endpoint creation

Your code is automatically containerised and deployed on Kubernetes

Available as SaaS or on-prem version

As you don't have to install and maintain kubernetes clusters, you give away some control of how this is done. You do have the option to set minimum and maximum amount of active instances

We support Python and R, but not any other languages

Specification

Specification Title	Specification Description
Compatibility	UbiOps serves and hosts any code (Python and R) as a micro-service with an API endpoint.
Automation	Every piece of code that you upload is automatically deployed to a kubernetes pod, making it scalable depending on your specifications.
Modular Design	Every piece of code is containerised and any dependencies are installed in a container too.
Monitoring	Using the integrations with various monitoring tools, you can measure accuracy, F1 scores, data drift and explain local predictions.
Training	Using the integrations with ML services, Sagemaker and other training tools, you can easily push your model from the training environment to production with Ubiops.
Business Efficiency	UbiOps speeds up the time to market of your data science projects.