Koverse Platform 3.0 Boosts Machine Learning for Big Data Pipelines in Production

Koverse has announced the 3.0 release of the Koverse Intelligent Solutions Platform (KISP), which includes significant advancements for data pipelines in production. KISP 3.0 supports thousands of machine learning (ML) jobs per day on complex data while still securing access at the data level, an important requirement in the wake of GDPR and other privacy regulations.

“KISP 3.0 provides access to state-of-the-art machine learning technologies while still supporting security, scale, and performance,” said Paul Brown, Koverse Co-founder and Chief Product Officer. “The platform is supporting complex data pipelines with thousands of transforms per day.”

KISP 3.0 is available immediately for customer implementation. Key capabilities include:

  • Transform SDK: The new Transform SDK makes it quick and easy to create reusable analytics that read and write user-defined objects. With the SDK, users can access new features in Spark 2 such as typed datasets, improved SQL capabilities, and ML model persistence.
  • Cluster Automation API: The new API allows developers to automate the creation, execution, and tear down of entire data flows from a client. This functionality makes it possible to orchestrate activities such as on-demand spinning up new instances of data flows and running complex data processing experiments.
  • Advanced Column-Level Security Controls: For a while, Koverse has supported obfuscating values in specific columns to protect sensitive data. Now users can search on encrypted or otherwise reversibly obfuscated values within a column and get matching results back, all while continuing to protect the original values from disclosure. This functionality makes it much easier to comply with data privacy standards such as GDPR and HIPAA.
  • Improved Data Processing Job Handling – KISP 3.0 includes improvements to how data processing jobs are handled, enabling instances of KISP to manage tens of thousands of data processing jobs including everything from many multi-stage ETL pipelines to repeated machine learning training and inference jobs.
  • Spark 2 Integration: KISP 3.0 streamlines access to more built-in ML algorithms in Spark 2 and allows machine learning vectors to be stored natively in Koverse datasets.
  • Hadoop 3 Integration: Organizations are able to future-proof their applications with the latest and greatest version of Hadoop.

“KISP 3.0 delivers multiple features that enable our customers to tackle the unique challenges that arise when bringing sophisticated, intelligent applications to production at scale,” said Aaron Cordova, Koverse Co-founder and CTO.

“Our customers are already using machine learning with complex data to unlock new revenue opportunities and competitive advantages,” said Jon Matsuo, Koverse President and CEO. “KISP 3.0 sets the stage for next level intelligent solutions in production.

About Koverse
Koverse powers intelligent solutions with complex data. Our technology is trusted by industry leaders in finance, biopharma, professional services, government, and technology. Our founders are former NSA data scientists who created Apache Accumulo, the fast and scalable data store with cell-level security. Accumulo is the foundation for the Koverse Intelligent Solutions Platform (KISP) which supports real-time indexing, machine learning, and precision search across disparate data. Learn how Koverse technology is powering PwC’s Know Your Vendor and many other intelligent solutions.