What if the next great cosmic discovery wasn’t made by NASA or a world-class astrophysicist—but by a teenager doing homework after class?

In a world where most 18-year-olds are worried about college applications or weekend plans, Matteo Paz was quietly scanning the sky—unearthing 1.5 million previously unknown objects in space. He didn’t use a telescope. He used code. And a machine-learning model he built himself.

This wasn’t a science fiction plot or a classroom exercise. It was real. So real, in fact, that it earned him $250,000, a peer-reviewed scientific paper, and a permanent place in the annals of space research. But the most astonishing part of this story isn’t the technology or even the prize. It’s what it reveals about the potential hidden inside each of us—and how, with the right mix of curiosity, mentorship, and perseverance, even a high school student can rewrite the map of the universe.

Because sometimes, the stars don’t just align. They wait to be found.

How a Stargazing Kid Became a Space Pioneer

Matteo Paz’s story doesn’t begin with artificial intelligence or astrophysics—it begins with awe. As a child, his mother would take him to public stargazing lectures at Caltech. These weren’t just family outings; they were ignition points. In those moments under the stars, surrounded by scientists and dreamers, something clicked. He wasn’t just looking at constellations—he was asking questions. Big ones.

By the time most kids were just starting algebra, Matteo was finishing AP Calculus BC in eighth grade through Pasadena Unified’s Math Academy. But his brilliance wasn’t just academic—it was driven by deep curiosity and a quiet confidence that he could contribute, even in a field dominated by PhDs and satellites. That mindset led him to the Planet Finder Academy in 2022 and then to Caltech’s Summer Research Connection program, where local high school students are paired with real research mentors.

Image Credits: Website @SocietyForScience

Enter Dr. Davy Kirkpatrick—a seasoned astronomer who had once been just like Matteo, growing up in a small town in Tennessee with a dream far bigger than his zip code. He saw something familiar in the young student: potential, raw and rare. When Matteo said he wanted to publish a paper—a goal far beyond the scope of the summer program—Kirkpatrick didn’t laugh, didn’t deflect. He simply said, “OK, let’s talk about that.”

That moment of belief—simple, supportive, and profound—became a launchpad.

From there, Matteo wasn’t just learning about astronomy; he was beginning to shape it. The spark had caught fire. Not because he had all the answers, but because someone gave him the space to ask the right questions. And that’s where all pioneers start—not with certainty, but with wonder.

Using AI to Explore the Universe

Most people think of space exploration as something that happens in high-tech labs or billion-dollar missions. But Matteo Paz cracked open the universe from a computer terminal at Caltech, not with a rocket or telescope—but with an idea, some code, and an AI model called VARnet.

The universe, as it turns out, leaves behind a trail of data. NASA’s retired NEOWISE infrared telescope had spent over a decade scanning the sky, collecting nearly 200 billion rows of raw observational data. That data sat largely untouched when it came to identifying what scientists call variable objects—celestial phenomena that change over time. We’re talking about blinking quasars, flickering newborn stars, eclipsing binaries, and other cosmic signals hiding in plain sight. The catch? There was too much data for any human to sift through manually.

That’s where Matteo saw a different path. Drawing on his passion for machine learning and his foundation in advanced mathematics—skills honed through years of intensive study—he designed an AI-powered system to detect the subtle rhythms in this massive dataset. The model didn’t just look at brightness or noise; it listened for patterns. Like a conductor picking up on changes in a symphony, his system could sense the faintest shifts in light curves, the heartbeat of the cosmos.

He combined wavelet transforms to reduce noise, and introduced a novel technique he called the Finite-Embedding Fourier Transform (FEFT) to extract recurring signals hidden in the chaos. Then, with deep learning layered on top, the system became fast and precise—processing each light curve in under a millisecond.

What did all that get him? A list of 1.9 million variable objects—1.5 million of which had never been seen before.

Among these were potential supernovae, supermassive black holes, and even phenomena that defied immediate classification. Some of the objects he identified may end up rewriting what we know about the life cycles of stars or the structure of galaxies.

But beyond the science, what Matteo accomplished was something more profound: he turned forgotten data into discovery. He showed that AI isn’t just a tool for industry or entertainment—it can be a telescope of the mind, a key to unlocking secrets the universe has been whispering all along.

Why Belief in Someone Can Launch a Legacy

Behind every young person who dares to reach for the stars, there’s often someone quietly holding the ladder.

For Matteo Paz, that person was Dr. Davy Kirkpatrick—a veteran astronomer at Caltech’s Infrared Processing and Analysis Center (IPAC) who, years earlier, had been a dream-filled teenager himself. Growing up in rural Tennessee, it was a high school science teacher who told Kirkpatrick and his mother that he had potential, offering a roadmap for his future. That act of recognition stayed with him. And when he met Matteo, he passed that torch forward.

In a world quick to doubt the inexperienced, Kirkpatrick did the opposite. He listened. When Matteo said he wanted to turn his summer project into a full-fledged research paper—a goal that could intimidate even graduate students—Kirkpatrick didn’t say, “That’s not realistic.” He said, “OK, let’s talk about that.” It was a small moment, but it echoed loudly in Matteo’s journey.

Their meetings became more than academic check-ins. They were conversations, sometimes 10% research, 90% deep dialogue about science, purpose, and life. And that connection, that genuine investment in Matteo’s growth, made the difference. Kirkpatrick didn’t just share data—he shared belief.

That belief multiplied. Kirkpatrick brought Matteo into contact with other experts at Caltech—Shoubaneh Hemmati, Daniel Masters, Ashish Mahabal, Matthew Graham—creating a network of support that turned a student into a collaborator. By 2024, Matteo wasn’t just being mentored; he was mentoring others—guiding new high school students through the same programs that had launched his own breakthrough.

This is the quiet power of mentorship. It doesn’t just teach skills—it validates dreams. It’s how a kid with a passion becomes a scientist with a voice. It’s how potential becomes legacy.

The Real-Life Challenges Behind the Headlines

While coding breakthrough algorithms and analyzing cosmic data, Matteo was also navigating the very real chaos of life. Earlier that year, he and his family were forced to evacuate their home due to the Eaton Fire in Pasadena. For most people, displacement alone would be enough to derail any project—especially one involving terabytes of astronomical data and custom-built AI models. But Matteo pressed on.

Then came the technical hurdles. The NEOWISE dataset was massive—nearly 200 billion observations. The data was messy, noisy, and full of quirks. Early versions of his algorithm would break down. Code would fail. Patterns he thought he’d found would disappear on closer inspection. Even the machine learning models, powerful as they were, often behaved unpredictably when confronted with real-world cosmic complexity.

But instead of stepping back, Matteo leaned in. He didn’t chase perfection—he chased progress. When things broke, he refined them. When the answers weren’t clear, he asked better questions. What might’ve been frustration became fascination. As he later said, “What’s exciting is that some variable phenomena I’m detecting don’t have obvious explanations. These quirky, mysterious objects are exactly those that spark new science and physics.”

That mindset—curiosity over comfort, questions over certainty—is what separates the dreamers from the doers.

It’s important to remember: breakthroughs are rarely born in perfect conditions. They rise through adversity. They take shape in cluttered labs, late-night coding sessions, and, sometimes, the chaos of evacuation orders and smoke-filled skies. Matteo’s success wasn’t just a testament to intelligence—it was a reflection of resolve.

Why This Discovery Matters for All of Us

On the surface, his project was a technical triumph: over 1.5 million previously unknown space objects identified through artificial intelligence, drawn from an ocean of forgotten NASA data. But zoom out, and the impact goes far deeper. Matteo transformed a dormant archive—nearly 200 billion data points collecting virtual dust—into a living, breathing map of the cosmos. With his model, astronomers can now revisit the NEOWISE dataset and uncover new clues about supernovae, black holes, stellar evolution, and perhaps phenomena we haven’t even named yet.

That’s science. But the story Matteo tells—through his action, not just his algorithm—is even more universal.

He reminds us that innovation often comes not from institutions, but from individuals willing to look at something old with new eyes. That age is not a barrier to impact. That AI, when guided by curiosity and ethics, can be more than a buzzword—it can be a bridge between what we know and what we’ve yet to imagine. And that mentorship, access, and belief aren’t side notes to success—they’re the scaffolding that makes it possible.

The ripple effects of Matteo’s work won’t end in a publication or a science fair win. His AI model, adaptable and time-sensitive, could be applied far beyond space: from analyzing stock market trends to tracking climate patterns or even detecting anomalies in public health data. In the right hands, tools like his don’t just chart stars—they chart the future.

And maybe the greatest lesson of all is this: the unknown is not a void to fear. It’s a canvas. Whether you’re a scientist, a student, or someone simply trying to find your way in the world, Matteo’s journey offers a quiet, powerful invitation—

To explore.
To ask.
To believe that what we do—what you do—can echo far beyond what’s visible right now.

Featured Image from Website @SocietyForScience