RocketML HPC Machine Learning & Science Platform

RocketML is a high performance compute technology based machine learning product that supports end to end machine learning steps. With its fast back end, customers can speed up model building, data preparation, model deployment steps at lowest total cost and at maximum value.

Packaged as a PaaS solution on Azure, RocketML HPC product can be deployed into customers private networks on Cloud and customized for their precise security requirements. Customers' data never leaves their private network.

Great models are built when Scientists, Engineers iterate on data, probe their problem space repeatedly. This data driven experimentation is time consuming on normal data science tools making them unproductive. On RocketML HPC platform, results appear at near real time when appropriate compute hardware resources are used. RocketML makes it easy to change cluster sizes on the fly without having to change screens or context. Furthermore, RocketML is hardware agnostic and supports both GPU and CPU combinations without any memory limits.

Entire HPC and Cloud complexities are abstracted, enabling Scientists and Researchers to focus on solving Science and Engineering problems be it in Seismic Image interpretation or Log data while drilling or Risk mitigation modeling or MRI analysis. The platform overcomes GPU memory limitations, batch size limitations and enables Deep Learning at scale.

Our innovations cut across the systems stack: the memory subsystem, structuring parallel computation across GPUs and machines, and interconnects between GPUs and across machines.

With Support for Text, Tabular, Image and Video data types, plus support for wide range of data sources such as CSV files, SQL databases, Azure Blob Storage, Hadoop, Spark, Snowflake and more, RocketML is the most performant and versatile Machine Learning platform in the market.