Live at the Johns Hopkins Bloomberg Center, OPM Director Scott Kupor, NobleReach CEO Arun Gupta, Navy CTO Justin Fanelli and Johns Hopkins APL’s Christopher Watkins make the case for “skills over credentials” and a new on-ramp into public service.
WASHINGTON, D.C. — A new episode of the Washington AI Network Podcast, recorded before a live audience at the Johns Hopkins University Bloomberg Center Theater, featured an exclusive interview with U.S. Tech Force founder Scott Kupor, director of the Office of Personnel Management. This federal government initiative is aimed at rapidly expanding the government’s technical talent pipeline as AI and emerging technologies accelerate faster than the public sector can hire and adapt.

The conversation, moderated by Washington AI Network founder Tammy Haddad, also included Tech Force partner Arun Gupta, CEO of the NobleReach Foundation; Justin Fanelli, chief technology officer for the U.S. Navy; and Christopher Watkins, chief mission engineering and integration officer at the Johns Hopkins Applied Physics Laboratory.
From the Hopkins Bloomberg Center stage, Kupor framed the initiative as a response to two converging pressures: the speed of technological change and an impending demographic crunch across the federal workforce.
“The pace of technology is just going to continue to accelerate,” Kupor said, warning the government is “woefully under-prepared from a talent perspective.” He added, “Only 7% of the workforce in government is under the age of 30… if we do nothing, we have a pending problem.”
Kupor said early interest has surpassed expectations, with aspirations to scale the program quickly. “It’s been phenomenal, way better than expected,” he said, pointing to ambitions beyond the initial target. “1,000 is our initial target, but my aspirations… we should be doing 5,000… 10,000.”
A central design feature, Kupor emphasized, is reducing the “all-or-nothing” nature of government service—particularly for early-career technologists who may be wary of committing for decades. “I don’t want people to feel like they’re making a 40-year decision,” he said, adding that movement between public service and the private sector can be “healthy for both.”

A skills-first model — and a new public-service pitch
Kupor also highlighted a major shift in federal hiring: placing skills ahead of traditional credentials. “You don’t need to have a degree at Tech Force,” he said. “The requirement for the job is can you actually perform the skills that we need you to perform.”
Gupta, whose NobleReach Foundation has built a pipeline of early-career technologists into government roles, argued the bigger challenge is cultural: modernizing how public service is sold—and experienced.
“We used to sell government as a 30-year career,” Gupta said. “What we’re saying now is: come for a year or two, learn, serve, and see what’s possible… and whether you stay or leave, you become part of the public-service ecosystem.”
Gupta said firsthand experience is also a trust-building tool at a moment of deep skepticism about institutions. “Once you’re inside government, you humanize it,” he said. “You understand that there are good people with positive intent doing hard work — and that changes how trust gets rebuilt.”



AI at the edge: decision advantage, trust, and “exponentially better” performance
In the episode’s national security-focused segments, Fanelli and Watkins described how AI is already changing mission-critical operations—especially when deployed close to real-world problems.
Fanelli said the value is measurable in outcomes that matter in operational environments. “What it looks like is time on mission, increased lethality, increased survivability, and increased power projection,” he said, adding, “There are places where it’s not linearly better—it’s exponentially better.”
Watkins emphasized that AI’s promise is tied to reducing cognitive overload and enabling faster, higher-quality decisions in high-stakes settings. “Think about cognitive load of the operators,” he said. “It’s really about decision advantage at the end of the day—who can make the best decision fastest?”
Both stressed that adoption ultimately depends on trust. Watkins pointed to explainable systems as a requirement, not a luxury. “How do you trust what the algorithms are providing to you?” he asked, describing explainable AI as a way to show “evidence accrual—why it’s saying what it’s saying—so the operator can trust and promote it in the system.”
Fanelli argued the fastest learning happens when technologists are embedded closest to the mission. “If you are closer to the problem and closer to the people with the problems, you’re going to learn more,” he said. “You’re going to figure out what applications are there.”The episode is available on YouTube, Apple Podcasts, Spotify and major podcast platforms, with a full transcript provided by the Washington AI Network. Here’s how the event was covered by NextGov.






























You must be logged in to post a comment.