1. Pipeline builder: New features have been added to the pipeline builder to give you more freedom and control over your pipelines.

  • A pipeline run can be exposed as a public REST endpoint so that you can integrate the pipeline to our application!
  • Additional pipeline validations have been added
  • Searching for a Block in the pipeline builder has improved with more information
  • Many new pre-built blocks have been added for use in the visual builder--ML models, data processing blocks, data source readers, data source writers, and more!

2. SDK/Jupyter Lab: The new SDK makes creating & testing blocks and pipelines easier

  • The SDK has been updated to make building blocks faster and easier. Also, more pythonic APIs for fine-grained control.
  • Pyspark is now included in RazorThink Platform so you can lever its power on big data.
  • You can store all trained models in the new model inventory using the SDK. The interface is coming soon!
  • The cookbook widget within Jupyter lab provides you all the SDK API documentation and quick guides you will need to complete your work.
  • The pipeline run widget will display the block metrics, separate and apart from the existing logs and other details.

3. Administration: Datasource offerings and improvements

  • You can add Oracle as a data source and use pre-built blocks to connect to your data
  • Under the libraries section, you can now search and sort both the User and System Libraries.
  • Engine Monitoring - are a data scientist, or in IT Operations, trying to figure out who is using up the servers? Built-in engine monitoring gives you transparency into all the pipelines that are running on all the engines!
  • Engine Management - You can see who did what and when with your Engine. For example, you can now get visibility to who paused your Engine.