About the conference

…something happened. Nobody classified it. Until now. Somewhere in this data, something is waiting to be found.

A full-day data science hackathon using real NASA and ESA space data. Participants work in teams to solve classification and discovery challenges using the same machine learning algorithms used every day in fraud detection, credit scoring, and fintech. Two tracks, two sides of the Force — one shared dataset from the stars.

The Two Tracks

Track 1: "The Data Strikes Back" Classify unidentified gamma-ray sources from NASA's Fermi-LAT catalog. Is it a blazar? A pulsar? A black hole candidate? Teams build machine learning models to classify what nobody has classified yet. Pure data science, epic dataset.

Track 2: "May the Forest be with you" Hunt for unreported transients in NASA and ESA sky survey data. Cross-match catalogs, find anomalies, and if a team finds something real — we submit it to the IAU Transient Name Server (wis-tns.org). That's the official international registry for astronomical discoveries, open to everyone. Their names, on an actual IAU-designated discovery. For real.

Dark Side vs. Light Side

Within each track, teams are split into two sides:

THE DARK SIDE
AI-assisted / vibe coding Use every tool. ChatGPT, Claude, Copilot — anything goes. Raw power, shortcuts, speed. The Sith approach: results at any cost.

THE LIGHT SIDE
Traditional coding No AI assistance. Pure skill, discipline, understanding. You earn every line of code. The Jedi way: knowledge through practice. At the end of the day, both sides present their results. Did the AI-assisted teams build a better classifier? Or did the traditional coders understand the data deeper and get more accurate results? Same challenge, same data, different approach. The comparison is the experiment.

Participant Requirements

Dark Side teams (AI-assisted coding)

• No coding experience required
• If you can prompt an AI, you can participate
• Bring your own laptop with access to your preferred AI tool (ChatGPT, Claude, Copilot, etc.)
• Must be able to explain the results at the end — understanding matters, not just output

Light Side teams (traditional coding)

• Basic Python knowledge required (functional, not expert)
• Familiarity with Jupyter notebooks is a plus
• Bring your own laptop with Python installed

Zero astronomy knowledge needed for either side. The host covers all the science context in the intro