Fearless at Everstar: A Former Regulator's POV

At Christmas, one of my closest friends gave me a bracelet with one word: FEARLESS. It was meant as encouragement. I didn’t know at the time it would become the theme of my year. Six months ago, I left the U.S. Nuclear Regulatory Commission to join Everstar, a startup building AI solutions to accelerate nuclear energy. “Fearless” is not a word I would have used to describe my government service. But it’s the word that has defined this transition.

A Different First Day

On my first day at Everstar, my boss told me: “Unless it’s an irreversible decision, just do it.” At the NRC, I would have paused, drafted, and asked for three approvals before moving forward. That one line told me life had changed

I had spent my career at the NRC, and I loved it. The NRC’s mission—protecting people and the environment—gave clarity to every project. It made recruiting easy, and it gave me purpose. But even in an organization built on mission, I started to see the limits.

The Regulator’s Lens

Rulemaking was the clearest example. We had strong ideas and worked hard to bring them to life, but it was always “speed up to slow down.” Writing was another sore spot. Every document began from scratch or with copy-and-paste, which meant mistakes. At one public meeting, the slides carried the wrong plant name. The rigor was there, but the tools weren’t.

I was fortunate to work across almost everything the NRC regulates: new reactors, operating reactors, and radioactive materials. Few staff had that range. I led big projects like developing the fusion regulatory framework. I was on task forces that tackled really hard problems, some of which totally changed our practices. The MO: focus on what matters most, and never lose sight of the larger system.

Choosing Something Different

I didn’t leave because I lost faith in the NRC. I left because I believed I could apply what I had learned in a new way. The federal government itself was encouraging people to seek higher-impact options. If agencies could experiment, why couldn’t I?

The job posting that caught my eye didn’t even name the company. It asked a simple question: what if licensing and regulatory work could be done more efficiently? That was enough to pull me in. Behind it was Everstar, using AI to cut through the slowest parts of nuclear without touching the margins that matter.

Culture Shock and Carryover

The culture shift was sharp. At NRC I lived in a permission system. At Everstar, the bias is for action. That disorientation faded into freedom once I realized how much progress small teams can make.

Some instincts carried over well. A safety-first, risk-informed mindset helps me decide what matters most. We can’t do everything, so we focus where the impact is largest. Other instincts I had to leave behind. At NRC I defended the agency in every conversation and weighed each word. At Everstar, people want candor. They want me to speak plainly, even when the truth is uncomfortable.

Early Signs of Impact

In six months, we’ve shifted from building infrastructure to shipping solutions. We now deliver sources-cited research, change-control evaluations, and complex licensing documentation. These aren’t academic tools; they save real time on projects that cost tens of millions.

The response has been encouraging. We’ve already finished work that delivered results for early customers. We’re getting inbound interest from major players. After my public talks, people often move from cautious to curious to committed. That progression is one of the clearest signs we’re addressing real needs.

What I’ve Learned About AI

AI isn’t magic. Without guardrails it can hallucinate, so we spend time managing memory, breaking up problems, and curating the right data. Garbage in, garbage out.

What surprised me most is that it can reason, not just summarize. Done right, it delivers believable results with sources cited. That matters in nuclear, where decades of knowledge sit in scanned or even paper-only documents. Unlocking that is the real accelerator.

The lesson is not to expect perfect answers in one shot. You iterate and refine. In this world, AI gives you first drafts fast and research right away.

Looking Forward

When Everstar succeeds at scale, we’ll slash the costs of nuclear licensing, construction, and operations by an order of magnitude. That won’t make us popular with everyone, especially those whose business depends on today’s costs. But it could reshape how nuclear energy is deployed.

For me, the next six months are about impact: turning pilots and enthusiasm into durable results. And for my former colleagues at the NRC, I hope the message is clear: I didn’t leave the mission behind. I carried it with me, and I’m working on it in a different way. I hope you’ll be inspired to operate this way inside the NRC—or to join me.