CelestiAI.vip was collaboratively developed with my hackathon team during the NASA Space Apps Challenge 2025 (Auckland), where we placed 2nd overall. As Team Lead and ML Engineer, I led the team, allocated tasks, made key project decisions, and designed the exoplanet detection pipeline — including training the Random Forest model and the habitability classifier. I worked closely with the team to integrate these components into a live web deployment, ensuring a functional end-to-end prototype.
| Name | Role & Contributions |
|---|---|
| Aayush Prakash | Team Lead / ML Engineer — led the team, allocated tasks, made project decisions, designed pipeline, trained Random Forest model and habitability classifier |
| Mingming Zhang | ML Support — cleaned code, preprocessed data, assisted with model training |
| Jonty McBreen-Graham | Backend, Git & Docker — managed Git workflow, handled technical Git issues, built Flask API bridging back & frontends and built & deployed docker container |
| Dmitrii Gomazenkov | Frontend & Visualization — developed UI using Three.js, implemented interactive 3D visualizations |
- Upload .csv telescope data
- Redirect users to a new page for results
- Process data using an ML model
- Display parsed data in 3D format using Three.js
Optional / Future Features:
- Community board for user results
- Ever-growing 3D map of confirmed exoplanets
Python, Flask, scikit-learn, pandas, numpy, Three.js, Docker
Docker (recommended):
git clone https://github.com/CelesteiAI/nasa-hackathon-2025.git
cd nasa-hackathon-2025
docker build -t nasa-hackathon-2025 .
docker run -d -p 8080 --name celesteiai-nasa-app nasa-hackathon-2025nasa-hackathon-2025/
├─ src/
│ ├─ app.py
│ ├─ static/
│ │ ├─ css/style.css
│ │ └─ js/base.js
│ └─ templates/
│ └─ index.html
├─ trained-models/
├─ dockerfile
├─ requirements.txt
└─ README.md
Leading this project as a first-year Computer Science student was an incredible experience. Together, we built a fully functional web app with live deployment.
Key learnings:
- Leadership: managing tasks and guiding the team under time pressure
- Machine Learning: designing and implementing algorithms for real-world data
- Collaboration: working across specialized roles (backend, frontend, Git)
- Execution & Usability: balancing technical depth with user experience
CelestiAI.vip demonstrates the power of collaboration, technical skill, and leadership, while making exoplanet science accessible to everyone.
Built with ❤️ for NASA Hackathon 2025