A new technology stack is offering government agencies a way to accelerate and simplify their AI data pipelines in the hybrid, multi-cloud environment.
During the 2023 NVIDIA GTC Conference, Government Acquisitions, Inc. (GAI) leaders previewed the new solution, developed in partnership with NVIDIA, Dell Technologies, Weka:IO, and Run:AI – that helps government agencies simplify their AI data pipelines.
“We’ve partnered with the best in the industry … to bring a holistic approach to how you can manage your data, no matter where it resides, and get that data over to your project when it’s needed,” said Mike Broadwater, vice president of technology services at GAI.
Broadwater explained how the Weka:IO solution solves storage platform challenges, accelerating and automating Machine Learning Operations (MLOps). He said the team started building the solution by looking at where data may reside – whether on-prem or in the cloud – and looked for a solution that could “scale across both of those mediums.”
“If it’s in your data center, you want the same storage view that you have in the cloud,” he said. “That reduces the administrative burden for the team that’s supporting you, as well as it presents a common interface for data fabric.”
As for the Run:AI portion, he said it helps agencies efficiently use hardware for machine learning workloads by giving agencies the power to manage compute resources.
“We can take and now manage all of your workloads, no matter where you have them,” Broadwater said. “This is where that interconnected data fabric becomes really important.”
The “secret sauce” of it all is the control plane that’s provided by Weka:IO, which he said helps bring everything into one view, and one data plane to work from and manage.
“Each new AI stack operates exactly the same way, with the same control plane, with the same tools and capabilities in any platform of your choice,” said Prem Jadhwani, chief technology officer (CTO) at GAI.
“Together, these capabilities significantly enhance and make your data science pipeline more efficient, and improves the productivity of your data science team and simplifies the deployment,” he added.