Projects
Full-Stack Portfolio Website with Live Analytics & Secure Deployment
Production-ready Flask application with real-time visitor tracking, cloud deployment, and backend security hardening
Project Summary:
- Designed and developed a complete full-stack portfolio website from scratch using Python, Flask, and Bootstrap.
- Engineered and deployed a live visitor tracking system featuring geospatial map integration to visualize global website traffic interactively.
- Implemented proactive security middleware to block malicious requests aimed at sensitive routes (e.g., .env, config files, AWS directories).
- Successfully deployed the application to a production environment (Railway) integrating a robust MySQL database architecture.
- Established a modular backend featuring RESTful endpoints, Jinja2 template inheritance, and an admin dashboard for content management.
Substructure Detection in Heterogeneous 3D Data Types
Unsupervised segmentation and substructure discovery across diverse volumetric and point-cloud datasets
Project Summary:
- Addressed the challenge of identifying hidden substructures in heterogeneous 3D datasets by developing a unified, data-driven analysis framework.
- Designed modular Python pipelines to process and analyze volumetric data, point clouds, and voxel-based representations originating from different experimental modalities.
- Applied unsupervised learning techniques, including clustering and density-based methods, to segment complex data without prior labels.
- Integrated statistical analysis and feature extraction to distinguish meaningful structural patterns from noise and acquisition artifacts.
- Enabled flexible experimentation across multiple data types, demonstrating adaptability of the approach to biological, materials science, and imaging datasets.
- Highlights expertise in scientific Python, 3D data processing, unsupervised machine learning, and algorithm development for real-world research data.
