Depending on an agency’s need, Nutanix may lean on partners such as NVIDIA to build the AI tools while it aligns with the IT department to deploy capabilities such as core AI infrastructure, as well as DevOps or Kubernetes expertise, in the field.
“When we bring those two components together with our existing partner relationships, we can make it that much easier to transition an idea or prototype into production, where it can be maintained by the IT team,” Langone says.
The next step is addressing what Langone calls “day 2 operations,” the work that must be done once an AI application is up and running. This can include implementing and optimizing AI models, troubleshooting models with poor performance or unexpected outputs, and building a scalable architecture that can grow with an agency’s needs.
EXPLORE: Federal chief AI officers are deploying flagship use cases of the technology.
Providing Granular Control Over AI Applications
Finally, agencies must enable more granular control over AI applications and closely monitor how they work. This is a key component of Nutanix GPT-in-a-Box 2.0. Whereas version 1.0 provided the integrated capabilities that let organizations run AI workloads in private cloud environments, version 2.0 enables central management of AI models.
“The 2.0 release was less of an evolution and more of a jump,” Langone says.
Now, an agency can deploy a model to different endpoints — human resources, procurement, a third-party contractor, a code-generation app — with differing controls for each business unit’s requirements.
Localized troubleshooting is among this option’s benefits, enabling IT teams solve problems more efficiently. Another benefit is the ability for agency leaders to identify usage patterns by line of business and see trends for resource allocation and budgeting purposes.
“Leaders can understand the value that AI brings to their organization, and they can set a path for growth,” Langone says.
UP NEXT: The State Department is developing a federal playbook for responsible AI deployment.
Accompanying version 2.0’s release was the launch of the Nutanix AI Partner Program, through which other vendors will help Nutanix users build, run, manage and secure generative AI apps atop the Nutanix Cloud Platform.
The partner program is important for the company’s federal customers because it gives agencies access to vendors that have been vetted based on their value proposition to the government, Langone says. Criteria evaluated include a vendor’s embrace of private cloud and its ability to meet the requirements of conversational document search, which is a key point of emphasis for many agencies.
“Ultimately, someone’s going to build an AI capability that solves a specific challenge for an agency, and they will need to run it somewhere,” Langone says. “That’s why it’s critical for agencies to have a platform where they can build, run and troubleshoot AI applications.”
Brought to you by:
