AI/Machine Learning Innovation of the Year
AI represents an opportunity to turn unstructured data into insight; promising the ability to innovate faster and extend competitive advantage. Organisations of all sizes and from all sectors want to make the most of this opportunity. However, the road to implementing AI has often proved complex and costly.
Enterprises are calling out for the features and functionality required to execute increasingly complex AI initiatives, through any phase or scale. A new approach is required. One that eliminates the complexities that are holding many back. This is precisely why Pure Storage has worked on pioneering industry-first AI solutions to ensure businesses can truly become data-centric.
AIRI - An Industry-First
One common challenge Pure took note of was when organisations were spending too long building and testing infrastructure, rather than focusing on the data insights AI can deliver. In light of this Pure joined forces with NVIDIA to architect AIRI, the industry’s first integrated AI-ready infrastructure for deploying deep learning at scale.
Put simply, AIRI removes the infrastructure ‘headache’ that many have faced; eliminating shadow IT problems and enabling data scientists to focus on adding real value – and identifying insights that can help change the world. Whether it’s diagnosing and treating cancer with Paige, or accelerating AI projects across every sector with ElementAI, Pure is at the forefront – putting data to work.
AIRI is a purpose-built, fully integrated, AI stack solution. It is designed with data architects, scientists and business leaders in mind, breaking down the barriers so everyone can operationalise AI-at-scale.
It solves the biggest challenges by eliminating software and infrastructure complexities, in three innovative ways: its unified configuration delivers maximum performance; the unified data platform ensures AI workflows and pipelines are seamless; and optimised software means customers can hit the ground running rather than spending months testing and developing best practices.
Completely scalable, AIRI eliminates the challenges of building an AI data centre, neatly and compactly delivering all the necessary components in a small form-factor. Customers can explore what AI can really accomplish by empowering data architects and scientists to deliver time-to-insight at scale.
Set-up takes hours (as opposed to weeks or months) as the entire stack is configured and tested as a complete solution, avoiding intricate configuration and tuning - whether dealing with NVIDIA GPU Cloud (NGC), speeding metadata management with Pure Tools, or delivering an elastically scalable resource with Kubernetes. In addition, this means time-to-model development and insights can be achieved in just days.
An example of this in action is Paige – revolutionising clinical diagnosis and treatment in oncology with AI. While pathology is the cornerstone of most cancer diagnoses, most pathologic diagnoses rely on manual, subjective processes developed more than a century ago. This is where Pure Storage comes in.
“With access to one of the world’s largest tumour pathology archives, we needed the most advanced deep learning infrastructure available to quickly turn massive amounts of data into clinically-validated AI applications,” said Dr. Thomas Fuchs, Founder and Chief Science Officer, Paige. “The powerful combination of DGX-1 and FlashBlade accelerates our mission to catalyse the medical industry with AI.”
Making AIRI Accessible for Everyone
As of September 2019, for those that do not manage their own infrastructure, Pure Storage has made AIRI-Ready Datacentres available. Designed to help customers leverage AI for innovation, the program provides a wide range of services from managed IT to colocation, try and buy, managed services and true Software-as-a-Service built on AIRI.
For example, customers that want an owned AIRI system but aren’t equipped to support it on-site, partners like Equinix, the world’s largest IBX datacentre and colocation provider, are equipped to support any AI initiatives built on Pure technology.
Breaking down Barriers with AI Data Hub
Many organisations today are stuck with a siloed, traditional analytics infrastructure - making data difficult to access and work with. AI Data Hub is an end-to-end AI pipeline solution designed to break down these barriers and give enterprise customers what they need to design, develop and deploy AI quickly, and at scale.
AI Data Hub extends traditional analytics and provides more performance and security at a lower cost. Data scientists can expect major improvements in productivity, faster time-to-market, and more informed data-driven models.
From cleansing and tagging, through modelling, training and production, AI Data Hub streamlines deployment for an entire AI pipeline instead of a single phase, and eliminates risk of failure when scaling from proof-of-concept to production.
Companies like Element AI, a global provider of AI software, require rapid access to massive amounts of data through the entire lifecycle of a project and relies on Data Hub to help them.
“We leverage Pure Storage to help IT professionals understand how to bring AI science projects into the mainstream data centre and reduces overall complexity for our customers,” said Ludwig Gamache, Head of IT at Element AI. “Advanced enterprises are looking for an AI solution beyond what’s largely available today. Pure Storage delivers not only fast access, but secure and reliable throughput for our development pipeline.”
Today’s enterprise organisations face a significant set of challenges when it comes to AI deployment, the most critical of which is a complete overhaul in infrastructural capabilities. Enterprise organisations that have existed and done business one way for decades now find themselves working hard to build a business for the future. To truly compete going forward will require large-scale, multi-phase AI initiatives, and Pure has innovated in this space with these particular set of challenges in mind.