New features of UbiOps – October 2020

On the 8th of October 2020, we have released new functionality and we made improvements to UbiOps SaaS.

Here is an overview of what is new in UbiOps:

Scheduler for model & pipeline execution

We added the possibility to create a request schedule for models or pipelines in UbiOps. This way you can make periodic requests if you have a model or pipeline that needs to run at a scheduled interval.

This can be especially useful if your model or pipeline does not depend on direct input data, but instead makes a connection to an external data source. In brief, no external cronjob is required anymore to make an API call to trigger execution.

The request schedules can be configured through the WebApp, API, Client Libraries and Command Line Interface.

For more information on how to use Request Schedules, see our documentation:

See how can you create a Request Schedule in 3 steps.

Request Schedule
Request Schedule


Other changes & improvements:

  • We made numerous performance and reliability improvements related to infrastructure and data routing. This results in quicker deployments, a lower request latency and higher throughput rates for models and pipelines.

  • We improved the status of batch requests. It is now more clear when a batch request resulted in a failed model or pipeline call. In addition, batch requests will now make better use of all available model instances to improve concurrency.

  • As an organization admin, it is now possible to assign roles to users that have a pending invitation for your organization. This way the user can immediately start using the platform after he or she completes the registration process.

  • We improved automatic detection of the structure of your uploaded model ZIP file. The directory containing your model does not need to be named ‘model_package’ anymore.

  • Miscellaneous other improvements and bug fixes.

Enjoy the new functionality! Stay up to date with the developments on our blog. Read our documentation for more information on UbiOps and how to use it.


Is there anything that you are missing in UbiOps? We would love your feedback and suggestions: Go to our feature request form.

Latest news

Turn your AI & ML models into powerful services with UbiOps