A California high school student, Matteo Paz, accomplished a remarkable feat by using artificial intelligence to uncover 1.5 million previously unknown cosmic phenomena. The work, which began as a summer project, not only earned formal praise from NASA leadership in early 2026, but also became a peer-reviewed scientific paper. This achievement marks a turning point in space exploration, demonstrating the power of youthful curiosity combined with advanced AI tools.
Young talent from California surprises the scientific community
Pasadena High School student Matteo Paz joined Planet Finder Academy in summer 2022, a program designed to engage students in real-world astronomical challenges. Under the guidance of Caltech scientist Davy Kirkpatrick at the Infrared Processing and Analysis Center (IPAC), the young man began an analysis of a vast archive of NASA data. The goal was to map the cosmos in an innovative way.
The vast amount of information came from the space agency’s NEOWISE telescope, which since 2009 has collected a decade of all-sky infrared observations. This dataset featured nearly 200 billion lines of measurements, containing countless celestial objects and distant phenomena awaiting interpretation. Initially, the research team planned to manually study a small portion, but Paz had a bolder vision.
How a Summer Project Transformed Cosmic Data Analysis
Paz applied his strong background in theoretical mathematics, programming, and time series analysis to develop an automated algorithm capable of processing the massive file. In just six weeks, he built a machine learning system with the ability to identify dim and changing light sources. Such objects have a brightness that changes in a very subtle or unpredictable way, often going unnoticed by human eyes or standard software.
The model started showing surprising results almost immediately, according to Davy Kirkpatrick. As Paz continues to improve, the accuracy and excitement for discoveries has grown exponentially. The system was able to identify objects that twinkled, pulsated and disappeared, revealing signatures of quasars, binary stars and supernovae, significantly expanding knowledge about the universe.
The application of advanced techniques and the research team
Paz’s innovation used Fourier transforms and wavelet analysis, mathematical tools crucial for studying complex temporal signals. These techniques were essential for uncovering subtle variations in the infrared spectrum that NEOWISE sampling alone might have missed. Some objects changed so slowly or for such a brief period that they were impossible to detect with conventional methods.
The detection of these variables is essential for the study of rare transient phenomena and cataclysmic variables, which do not follow predictable cycles. During the process, Paz collaborated closely with Caltech researchers, including Shoubaneh Hemmati, Daniel Masters, Ashish Mahabal, and Matthew Graham. Together, they refined the system to analyze data from across the entire sky.
The catalog that drives observation at global observatories
The joint work resulted in the creation of a gigantic catalog containing more than 1.5 million variable light sources. This vast database was released publicly in late 2025 and is already being used to drive active observations at cutting-edge facilities such as the Vera C. Rubin Observatory, which began operating based on this data earlier this year.
The information collected and cataloged by Paz and his team is providing immediate insights into stellar evolution, the behavior of distant galaxies and high-energy cosmic processes. The young student’s contribution is now directly linked to the cutting edge of global astronomical research.
The democratization of space science through artificial intelligence
What Paz’s discovery demonstrates, in practice, is a seismic shift in modern astrophysics. Tools that once required institutional supercomputers and entire research departments are now accessible to high school students who, with a laptop and open source data, can decipher the universe. NASA publicly recognized the importance of Paz’s discovery in early 2026, validating that revolutionary AI-driven innovations can emerge from anywhere.
Matteo Paz developed his skills through the Pasadena Unified School District Mathematics Academy, a public program that nurtures mathematical talent. The tools he used, from algorithmic modeling to computational astrophysics, are typically reserved for graduate studies. This scenario highlights the democratization of knowledge and research tools, allowing young and curious minds to contribute significantly to science.
The future of research: from high school to large telescopes
Paz’s momentum did not stop at astronomy. He hopes to expand his AI system to other areas, given that the algorithm can analyze any temporal data. This includes sectors such as finance, environmental monitoring or neuroscience, where subtle changes over time reveal critical patterns. The project highlights how tools that drive cosmic discoveries can also offer insights into complex systems on Earth.
Davy Kirkpatrick’s mentoring was crucial, demonstrating a commitment to supporting students’ potential. The true scale of the revolution that Peace’s approach represents was confirmed in February 2026, when the Vera C. Rubin Observatory in Chile activated its real-time alert system.
- On a single night, February 24, the system issued nearly 800,000 alerts to astronomers globally.
- The alerts signaled events such as exploding stars, variable stars, active black holes and asteroids.
- When the ten-year Legacy Survey of Space and Time begins, Rubin is expected to issue up to seven million alerts per night.
This avalanche of data is impossible for any human team to analyze. The work is delegated to machine learning systems, which, distributed by automated “intermediaries”, decide in seconds which flashes of light deserve further analysis. A teenager’s pioneering methodology thus became the only way to do modern astronomy.
