A software developer says he used Anthropic’s Claude Code and OpenAI’s Codex to identify a possible exoplanet in publicly available NASA telescope data, prompting the space agency to schedule follow-up observations. The candidate remains unconfirmed, but the findings suggest a repeating signal from a star about 116 light-years away that could be consistent with a planet passing in front of it.
Developer Pavel Rabtsevich identified the pattern while examining observations collected by NASA’s Transiting Exoplanet Survey Satellite, or TESS. The signal appeared in 2025 data and was subsequently traced to separate observations from 2020 and 2018.
The star appears to dim by approximately 0.05% for about two hours every 3.18 days. If the dimming is caused by a planet crossing in front of the star, the possible planet would be around 1.4 times the size of Earth.
FCRF Launches CP-FRM to Build India’s Next Generation of Fraud Risk Professionals
Signal Traced Across Multiple Years of TESS Data
Rabtsevich’s investigation began with the recurring decrease in the star’s brightness. Such changes can be associated with planetary transits, when an orbiting planet temporarily blocks a small portion of its host star’s light.
Finding a repeating signal, however, does not establish the existence of a planet. Similar patterns can be produced by nearby stars, instrumental problems or other astronomical phenomena.
Rabtsevich said he used AI agents to examine alternative explanations and subject the candidate to a series of checks. The work included testing whether nearby stars, telescope glitches or other effects could account for the observed dimming.
He spent about two weeks examining the candidate, conducting 74 analyses and writing more than 1,000 scripts during the process.
Claude Code Helped Process and Test the Candidate
Claude Code was used to process TESS observations, develop analysis scripts, fit potential planetary transits and test competing explanations for the signal.
OpenAI’s Codex and separate AI sessions were also used to review parts of the work. The tools assisted with tasks that would ordinarily require substantial programming effort, including processing astronomical datasets and repeatedly testing different hypotheses.
The project illustrates how coding assistants can be applied beyond conventional software development. Tools originally designed to help developers write, debug and test code can also be used to automate parts of scientific data analysis.
In this case, Rabtsevich applied AI coding agents to publicly available telescope observations rather than relying on them simply to generate software. The agents helped him explore historical datasets and determine whether the suspected signal appeared consistently across different years.
NASA Follow-Up Will Test Whether Candidate Is a Planet
The possible exoplanet has not yet been confirmed. Further observations are required to determine whether the recurring dimming is genuinely caused by a planet or by another source.
NASA has scheduled follow-up observations to test Rabtsevich’s findings. Those observations could provide additional evidence about the nature of the signal and whether it behaves in a manner consistent with an orbiting planet.
The case is notable because the analysis was carried out by a developer rather than a professional astronomer. It demonstrates how AI-assisted programming tools may make complex scientific datasets more accessible to individuals outside traditional research institutions.
For now, the object remains a candidate. Confirmation will depend on further observations and whether the signal withstands additional scientific scrutiny.
About the author — Suvedita Nath is a science student with a growing interest in cybercrime and digital safety. She writes on online activity, cyber threats, and technology-driven risks. Her work focuses on clarity, accuracy, and public awareness.
Follow for daily updates on cybercrime, corporate fraud, DFIR, hacking, investigations, and digital forensics
