The Naval Postgraduate School (NPS) is building an artificial intelligence (AI)-ready force the way a teaching hospital trains clinicians: rigorous classroom learning, side-by-side practice on world-class technology, and real problems. Through its Artificial Intelligence Task Force (AITF), Digital Trident AI Challenge, and academic and leadership programs, NPS pairs tiered AI education with secure, scalable compute and operator-in-the-loop experimentation so students and faculty can turn promising ideas into useful capabilities that commands can adopt.
Education that sticks and scales
NPS’s education model is deliberately tiered: foundational concepts, immersive labs, advanced developer coursework, and executive-level learning for decision-makers.
Hands-on work is built into the NPS curriculum. “To truly understand machine learning algorithms – their purposes, capabilities, and limitations – you have to use them,” said Mathias Kolsch, an associate professor of computer science at NPS. “Most of our courses at NPS have a lab component for hands-on work that reinforces the lesson content.”
Executive offerings include a three-day course, “Leading Data- and AI-enabled Organizations,” developed with the Department of Defense Chief Digital and AI Office. Since early 2022, about 400 leaders have taken the course, and NPS has received strong feedback on its national security context, Kolsch said.
Beyond the classroom, executive students work through mission-relevant breakouts to connect lessons to real hurdles back at their organizations, an exchange that Kolsch says “has proven really, really effective” for accelerating adoption.
NPS’s collaborative approach extends to the composition of student groups. Public sector personnel, international allies, and industry sit shoulder to shoulder. “There’s nothing like talking to the end user to get the answer. Especially when that person sat next to you in class last quarter,” noted retired U.S. Marine Corps Col. Randy Pugh, vice provost for warfare studies and director of the AITF at NPS.
A working lab for the fleet
The AITF is “NPS’ engine for accelerating AI readiness across the Fleet [and] supporting establishment,” integrating advanced education, state-of-the-art infrastructure, and applied research and prototype development to transform concepts into operational capabilities, according to NPS.
Partnerships are central to the AI acceleration effort. NPS and NVIDIA operate under a Cooperative Research and Development Agreement that brings operators, NPS researchers, and NVIDIA experts together to develop AI solutions that address capability gaps in complex environments. The NVIDIA AI Tech Center at NPS, one of only two in the United States, provides a shared, hands-on space with modern hardware and platforms like NVIDIA AI Enterprise with NIM and NeMo microservices and NVIDIA Omniverse to support collaboration and prototyping of solutions ranging from decision support to predictive maintenance and digital twins. Through rapid iteration, real-world data, and domain-informed experimentation, these efforts shorten the timeline from concept to capability, according to NPS.
Compute that meets the mission
Education and solution development will not make it to mission implementation if the technology does not support real-world requirements. The NPS-NVIDIA partnership emphasizes deploying infrastructure and AI software across environments. The stack includes edge devices (e.g., NVIDIA Jetson), on-premises high-performance systems (e.g., NVIDIA DGX systems with NVIDIA Blackwell GPUs), and platforms like Omniverse for high-fidelity modeling and simulation, so teams can design, test, and scale AI technologies that meet the demands of complex operational environments.
In modeling and simulation courses and projects, for example, students can expand from AI experimentation on their computers to larger Dell Technologies clusters, each with four NVIDIA L40S GPUs, and, for the biggest problems, to an NVIDIA DGX GB300 rack-scale platform, said Perry McDowell, NPS’s lead for Omniverse and other modeling and simulation efforts.
The NVIDIA DGX GB300, for example, will support everything from complex mission planning to autonomous systems simulations to disaster recovery, NVIDIA said.
NVIDIA formalized its relationship with NPS in late 2024 and brought in Dell Technologies, Sterling, and an ecosystem of collaborators to ensure mission effectiveness, according to NPS. Dell and Sterling delivered the hardware, storage infrastructure, and systems integration required to deploy the Omniverse at NPS.
“The installation of the Omniverse instance at Naval Postgraduate School is a good example of cloud for many – but on premises for us,” Pugh said. “As we move into more sophisticated hardware, we are figuring out how we partition it across environments, from unclassified to secret and above, and how we scale it from the tactical edge to the enterprise infrastructure. Working with Dell Federal, a company that has done this for decades – they know the questions – and often the answers – before we’ve even asked.”
NPS is taking an AI factory approach to AI – implementing an end-to-end operational framework that will be backed by the Dell AI Factory with NVIDIA, a validated platform that includes everything from hardware to software to services, capable of running multiple AI workloads.
Sprints that ship
In August, seven NPS faculty-student teams were awarded a total of $825,000 in funding to develop AI solutions to critical public sector challenges through the Digital Trident AI Challenge, a collaborative program among NPS, the NPS Foundation, and NVIDIA. It leverages NPS relationships with additional industry partners, including Dell Technologies, that have contributed hardware, software, and expertise to NPS, Pugh said.
From 38 proposals, the seven teams were chosen for one-year research sprints based on technical merit, use of applied AI, interdisciplinary collaboration, operational relevance, and potential public sector impact. The goal is to accelerate AI adoption through applied experimentation in secure, real-world settings.
“All of the complex problems that we are trying to solve are interdisciplinary – requiring business and international relations professionals as well as scientists, industry experts, and operational specialists. That’s why Digital Trident AI Challenge teams must be interdisciplinary,” Pugh said. “You have to extend outside of your own area of expertise. And you have to go beyond research to deliver a capability in a timely manner so it can make a difference to the end user.”