NVIDIA and Pure Storage Power Ai for Oncology with PAIGE

Print Friendly, PDF & Email

The battle against cancer is now being fought at the data level. A new Ai Startup called PAIGE hopes to revolutionize clinical diagnosis and treatment in pathology and oncology through the use of artificial intelligence. Using Igneous Unstructured Data Management as-a-Service, PAIGE will protect and manage enormous datasets of anonymized tissue images for diagnosing and treating various forms of cancer.

Enabling researchers and scientists to directly access and use their data, and to collaborate with each other without going through an overburdened IT team—that’s the future of unstructured data management,” said Kiran Bhageshpur, CEO of Igneous. “The cutting-edge research and product development that PAIGE is doing necessitates cutting-edge data management and protection, and we are proud to support them in revolutionizing how to diagnose, treat, and prevent disease.”

This kind of work requires HPC technology. Through its partnership with Igneous, PAIGE will be able to securely and efficiently manage 8 petabytes of unstructured data, including anonymized tumor scan images and clinical notes, as part of an integrated machine learning-based healthcare AI workflow anchored by an industry-leading Pure Storage FlashBlade and NVIDIA GPU compute cluster.

Igneous provides the intelligent system of record for programmatically protecting PAIGE’s research-critical datasets used to train algorithms that assist in understanding, diagnosing and treating cancer. In addition, Igneous enables PAIGE’s data scientists to process, index, and move datasets as required for iterative model training and feature refinement. Igneous also serves as the long-term protection and retention tier for both the training datasets and results data post-processing.

PAIGE selected Igneous for its API-based object and file support with Pure FlashBlade, the ability to keep data at scale securely onsite in its native file format, as well as the Igneous managed as-a-Service model. The ability to use any file or object protocol provides PAIGE with crucial flexibility to experiment with new analysis technologies and to move quickly as a startup at the forefront of their field.

We selected Igneous because of its unique ability to organize and protect billions of images and its tight, API-based integration to our Pure Storage/NVIDIA compute cluster used to build our machine learning models,” said Ran Godrich, AI Research Engineer at PAIGE. “PAIGE’s mission is to revolutionize pathology through AI in order to fundamentally improve the efficacy and efficiency of clinical pathology diagnosis, while improving the care of patients around the world. Igneous engineers and experts are helping us realize our mission by developing our data infrastructure in the most secure way possible—not only in storage, but also the data management systems themselves. Their unwavering commitment to our success and the protection of our datasets is absolutely key to building a high-performance compute infrastructure.”

Sign up for our insideHPC Newsletter