Kinetica + NVIDIA RAPIDS Speed Up Predictive Data Analytics with the Power of GPUs

We’re very excited to announce the integration of Kinetica and RAPIDS!

Customers have been using the Kinetica Active Analytics Platform to do large-scale data preparation, as well as model inferencing and audit (square & circle, below).

With the integration with RAPIDS, we can now accelerate the whole end-to-end machine learning pipeline (including triangle, above).

Companies across industries like automotive, telco, oil and gas, retail, and financial services have been using Kinetica to explore their data at scale and to integrate machine learning models into active analytical applications. Now, with Kinetica and RAPIDS together, we enable the full end-to-end GPU-accelerated machine learning pipeline, streamlining workflows regardless of industry. 

Key capabilities of the Kinetica and RAPIDS integrated solution include:

•  Managed Jupyter Notebooks within the Kinetica Active Analytics Workbench, pre-integrated with RAPIDS. This is a fully managed, multi-user, GPU-accelerated data science environment.

• Interactive Data Exploration with Kinetica’s GPU-accelerated OLAP capabilities (that include tiered storage), enabling data scientists to explore the entire data corpus interactively. In cases where location analysis is required, data scientists can visualize data geospatially right within the Jupyter Notebook.

• Large-Scale Feature Transformations at scale, be it filters, joins, or complex geospatial functions. 

• Accelerated Training, where once feature transformations are complete, data can be seamlessly transferred to RAPIDS for training via a Python API call, right on the GPU.

• Model Deployment and GPU-accelerated Inference are automated with Kinetica, in continuous, on-demand, or batch modes. There’s no need to worry about deployment, network configuration, or scaling. Once deployed, Kinetica automatically orchestrates the full analytical pipeline – from ingest to database to model and back to database and downstream applications.

• Audit with Kinetica, which tracks the full data lineage, including raw data, feature transformations, and model output for all the model deployments. Our easy-to-use search tool provides an instant ability to do a full model audit or find a needle in the haystack for a specific inference.

These new machine learning capabilities help organizations develop more accurate models faster, and put them into production to get the benefit of the deep insights machine learning can bring. There are also features available in Kinetica to support governance, transparency, and repeatability. Kinetica’s integration with NVIDIA RAPIDS empowers data scientists and brings them into the center of modern, data-driven, enterprise-scale decision-making. 

To learn more about Kinetica + NVIDIA RAPIDS, read our data sheet and our press release.

For more on machine learning, check out our machine learning white paper.

This two minute video explains how you can build machine learning into your active analytical applications.

Irina Farooq is chief product officer at Kinetica. You can follow her on Twitter @IrinaFarooq.