The Capstone Project I'm developing in my College focuses on Real-Time threat detection using YOLO v11 and EfficientNetV2. This deep learning project aims to identify potential security risks through video analysis and identify threats in X-ray scans, with a specific focus on detecting weapons. Currently in development, this system will integrate with existing security infrastructure to provide automated alerts and monitoring capabilities (it's a long shot i know). To provide a comprehensive solution to the problem we have also designed an adaptive frontend that integrates these two models together.

Upon experimenting various CNN models like InceptionResnetV2 and Vision Transformers, optimizing them to identify anomalies in Dicom Slices and then deploying a Vision Language Model to describe what the anomaly is and in which slice it is presented. This project aims to assist radiologists in identifying and diagnosing anomalies more efficiently, reducing the time required for manual analysis and improving the accuracy of diagnoses. The final developed system will also provide a user-friendly interface for visualizing the results and integrating with existing medical imaging systems (hopefully).

Sometimes projects aren’t just about showing off skills—they’re about satisfying pure curiosity! I used this project as a hands-on experiment to finally understand client-side APIs. I was always confused about how they work, especially with async operations and client-side calls. So, I built simple projects that call three different API endpoints, which helped me see how requests, responses, and asynchronous flows all come together in real time. Big shoutout to GitHub Copilot, which totally transformed my coding game by whipping up CSS and streamlining my workflow(it's free labour). It’s crazy advanced and makes coding not only more efficient but also a lot more fun!

MedGrid is a comprehensive web-based Hospital Management Information System designed to streamline healthcare operations. Built with React and Material UI, it features secure authentication, patient and doctor management, room tracking, billing systems, and pathology report management. The application utilizes Supabase for backend services and database management, providing a responsive and intuitive interface for healthcare professionals. This project demonstrates full-stack development capabilities and modern web application architecture, with features like real-time data updates, comprehensive CRUD operations, and role-based access control.

A real-time speech-to-text web application that leverages the browser's native Web Speech API to convert spoken words into text. This project features voice command controls, real-time transcription, text editing capabilities, and export functionality. Built with vanilla JavaScript, it demonstrates proficiency in modern web APIs and provides an accessible, user-friendly interface for speech recognition. The application includes features like pause/resume functionality, text formatting options, and the ability to save transcriptions, making it a practical tool for note-taking and accessibility purposes.

This capstone project, a key milestone in my data science journey, showcases my ability to address real-world challenges through data-driven solutions. Using Python and essential libraries like NumPy, Pandas, and Scikit-learn, I analyzed a complex dataset to uncover actionable insights and develop predictive models. The project highlights my skills in data wrangling, exploratory data analysis, and machine learning, while reflecting my problem-solving abilities and attention to detail. This experience deepened my expertise in applying theoretical knowledge to practical scenarios, reinforcing my passion for leveraging data science to create impactful outcomes.

My attempt at developing an automated experience by building an expense tracking system using Google Forms, Google Sheets, and Make.com. The system collects user inputs using Google forms (I had to start somewhere), categorizes expenses, sends email notifications, and provides summaries through an email. I have a lot to do and this is an ongoing project. Check out the LinkedIn for more details and feel free to reach out if you have any questions or suggestions! This project showcases my ability to leverage no-code tools to create efficient and user-friendly solutions. By integrating different platforms, I automated a process that would typically require manual data entry and analysis, demonstrating my skills in automation and problem-solving.