How an AI Project by High School Students Helped NASA Discover 1.5 Million Space Objects

How a high school student’s machine learning project led to real-world impact and a $250,000 prize

High School Student Uses AI to Analyze NASA Telescope Data

Matteo Paz didn’t have access to a lab or telescope. What he had was curiosity, Python, and a public NASA dataset.

While most researchers were focused on tracking asteroids, Matteo trained an AI model called VARNet to detect everything else: variable stars, black holes, cosmic flickers that had gone unnoticed.

The result? His model flagged 1.9 million objects. Over 1.5 million were brand new to astronomy.

Machine Learning in Astronomy: How Matteo Did It at Age 17

He used data from NASA’s NEOWISE mission, which scanned the entire sky in infrared for more than a decade. That’s hundreds of billions of data points enough to make most heads spin.

But Matteo didn’t get overwhelmed. He built a machine learning model that analyzed brightness fluctuations in stars over time, identifying patterns using wavelet transforms and Fourier analysis.

This wasn’t just a science fair project. His work was published in The Astronomical Journal. And he took home $250,000 at the Regeneron Science Talent Search https://www.societyforscience.org/regeneron-sts/2025-student-finalists/matteo-paz/, one of the most prestigious science competitions for high schoolers in the U.S.

The Power of Mentorship in High School STEM Projects

Behind Matteo’s success was a key figure: Dr. J. Davy Kirkpatrick, senior scientist at Caltech. When Matteo told him he wanted to publish his research, Kirkpatrick didn’t say, “Maybe later.” He said, “Let’s talk about that.”

That one line made all the difference. It’s proof that real mentorship isn’t just helpful, it’s essential.

Why AI Projects for High School Students Matter More Than Ever

If you’re a high schooler or the parent of one, here’s what Matteo’s story proves:

• You don’t need a PhD to use real data. You need direction.

• AI and machine learning are not college-level-only tools anymore.

• High schoolers can publish. They can compete. They can get taken seriously.

• The right mentor can turn a vague interest into a world-class result.

From Curiosity to Code: Star Gazer, Built by a High School Student

A few months ago, a student at BetterMind Labs took on a challenge that felt way out of reach: tracking how star brightness shifts over time. Inspired by space missions like NEOWISE, the student built Star Gazer, a tool that uses machine learning to classify variable stars based on their light curves.

He didn’t start with a background in astrophysics. But he had questions and a mentor who helped her follow them to a working prototype. Her project didn’t just end with a presentation. He’s now exploring real-world datasets and refining them further for publication.

Projects like this are becoming more common, not because students are “geniuses,” but because the tools, guidance, and ambition finally line up.

Final Takeaway: The Future of STEM Belongs to Students Who Start Early

Matteo Paz didn’t wait for college. AI isn’t something to study later. It’s something to build with now, especially in high school.

And if you’re wondering whether students can really do this... they already are.

SEO-Optimized Headers Recap:

• Teen builds AI with NASA data

• High school student AI project

• Machine learning in astronomy

• AI research for high school students

• Mentorship in STEM education

• High school space research projects

• BetterMind Labs

• Astronomy projects for teens

• Regeneron Science Talent Search winner

• Student science competitions with real data