Modular Satellite Segmentation Pipeline

I built a modular deep learning pipeline using TensorFlow and SentinelHub to segment satellite imagery with a U-Net model. The system includes automated preprocessing steps like band selection, normalization and supports retraining on custom targets. It uses a lightweight classifier to prefilter input images and reduce false positives, with final outputs exported as KML for visualization.

View on GitHub

Personal Website

Built a full-stack web application using Python and Django. The app is deployed and hosted on an AWS EC2 instance, where I handled the setup and configuration of the server. This project involved building the front-end and back-end, creating a custom database, and ensuring the app was functional and accessible online.

View on GitHub

Bookworm

I collaborated with a team of four to develop and deploy a PostgreSQL database on AWS EC2, including over 433,000 ratings and 130,000 book records. I wrote and optimized SQL queries to ensure fast and efficient data retrieval across diverse book and user information. Additionally, I utilized Node.js and React.js for full-stack development, delivering a seamless and interactive user experience.

View on GitHub

Astropy, Contributor

Actively contributing to Astropy, an open-source Python library for astronomy and scientific computing. Working within a large collaborative codebase, I’ve engaged with real-world software engineering practices such as version control, code review, and automated testing. Through ongoing contributions, I’m building experience in open-source collaboration while deepening my skills working in large codebases.

View on GitHub