This Portfolio Site Built with Astro during my freetime is optimized for speed and search engines, designed with recruiters and tech enthusiasts in mind to efficiently access information about my skills, experience, and projects. 💡

This web app streamlines secondary school operations, offering an intuitive interface for assessment management and dynamic data visualization for parents. Integrating machine learning, it predicts student performance based on assignment scores and tests, empowering students. Efficient Administrative and Management systems enhance overall institution operations.

Empowering academic collaboration, this Django web app lets authorized university users upload and access past questions, projects, and textbooks. With features like secure login, role-based access control, signal-driven reports, and automatic thumbnail generation, the app promotes knowledge exchange and streamlines resource management through robust performance and detailed log management.

Translated a MATLAB-based Ki67 cell counting code into a Python library and developed an API backend. This backend, hosted on a server, seamlessly processes image inputs with user-defined thresholds, providing an accurate count of Ki67 cells. The deployment enhances accessibility to this critical research tool, showcasing my proficiency in biomedical informatics, algorithm translation, and API development for impactful scientific advancements..

Led the development of Roulette Aid, a desktop application using Python and PyQt, with a focus on machine learning for 90%+ accuracy in predicting roulette outcomes. Key achievements include implementing a secure authentication system, designing a flexible licensing system, and establishing update functionality through an FTP server. Efficient process management was ensured using signals and queues, while Code Obfuscation was applied to safeguard the codebase and intellectual property.

Developed robust computer vision models using MATLAB for a client project, leveraging advanced image segmentation techniques. The models were specifically designed to autonomously distinguish skin lesions from ground truth images, showcasing a meticulous approach to medical imaging. Evaluated the model's performance through meticulous quantification using dice similarity scores and strategically selected relevant features, ensuring a comprehensive and effective solution for dermatological image analysis.

Spearheaded the development of a dynamic job board application using Django, Jinja templating, and Bootstrap for frontend design. Leveraging Django ORM and Beautiful Soup, I automated job scraping from mulitple sources, enhancing the platform's efficiency. With streamlined resume submissions, user-star ratings, and intricate filtering options, the app provides an exceptional user experience. Prioritizing strong user authentication and meticulous error monitoring, I employed Celery and redis for task scheduling and comprehensive cron job monitoring. The result is a robust, reliable, and feature-rich job board platform.