By Gaurab Chhetri - Hackathon Project

Beyond Blue - An AI-powered platform to explore exoplanets

Beyond Blue is an AI-powered platform that predicts exoplanets using NASA data and lets you explore them in a social media–style feed with AI-generated stories and summaries.

Beyond Blue is an AI-powered platform that brings space exploration to everyone. Built during RiverHacks 2025, it lets users explore NASA’s exoplanet catalog1 through a sleek, social media-style interface enriched with AI-generated summaries and planet stories.

Instead of browsing spreadsheets of astrophysical data, users can experience each planet as a living narrative, complete with discovery context, classification, and AI insights.

🚀 Inspiration

The idea for Beyond Blue started with a simple question:
“What if discovering exoplanets felt as intuitive and engaging as scrolling a feed?”

Exoplanet datasets are fascinating but inaccessible to most people. Our goal was to humanize scientific data using storytelling, AI, and thoughtful design, making astronomy exploration interactive and fun.

🪐 Highlights

🌠 The RiverHacks Journey

We built Beyond Blue in 24 hours; from brainstorming to deployment. Each team member took on key roles across data processing, model training, and frontend design.

Despite sleepless coding sessions, we built a working prototype that blended machine learning, storytelling, and user experience into one cohesive platform.

🏅 Winner – Bold Innovation in Search & AI Track
Awarded $1,000 for advancing AI-powered data exploration.

RiverHacks Win Photo View the announcement →

⚙️ Tech Stack

LayerTechnologies
FrontendNext.js, React, Tailwind, Radix UI
BackendNode.js (Express), TypeScript
AI/MLFastAPI, scikit-learn (Random Forest)
DatasetNASA Exoplanet Archive

💬 Reflection

Beyond Blue reminded us that creativity thrives under pressure. In just one night, we transformed complex data into a human-centered AI experience, a small step toward making science feel personal and alive.

👨‍🚀 Team

Built with an incredible group of collaborators:
Amul Poudel2, Binod Dhakal3, Rastrabhusan Dahal (Hritik)4, Samar Ranjit5, Sudin Katuwal6.

References

Footnotes

  1. NASA Kepler Object of Interest Dataset

  2. https://www.linkedin.com/in/amulpoudel/

  3. https://www.linkedin.com/in/binod-dhakal/

  4. https://www.linkedin.com/in/rastrabhushan-dahal/

  5. https://www.linkedin.com/in/samarranjit/

  6. https://www.linkedin.com/in/sudin-katuwal-111926361/