Relevant Coursework

Engineered a cross platform application to fetch, analyze display market data for stocks and cryptocurrencies, using Flutter with python as backend, REST APIs for real-time market data integration & Firebase for database and secure authentication. Developed a stock price prediction model using Python’s machine learning libraries, TensorFlow and Keras, leveraging LSTM neural networks for time series forecasting

Developed a compiler in C++ that includes both a lexical analyzer and a parser-interpreter for the SFort95 language. The tool reads and interprets SFort95 programs according to defined grammar rules, evaluating them and generating accurate output. It also performs syntax validation and provides meaningful error messages to help identify parsing issues or invalid tokens during compilation.

Developed a full-stack data breach analysis tool using MERN stack within a 24-hour hackathon. The app uses MongoDB for data storage and Express.js/Node.js to serve secure REST APIs with JWT authentication, React Router and Material-UI for the UI and Recharts for interactive charts. Flask backend runs predictive models using Python libraries Scikit-learn, TensorFlow and Keras with results visualized through Matplotlib and Seaborn.

Designed a production-grade Python web scraping and data pipeline on Linux to monitor NJIT's CSOD Career Center for job listing changes. The script authenticates via SAML (UCID, password, Duo) leveraging Selenium and dynamic DOM traversal, simulates user navigation and scrapes job listings across all paginated search results using XPath. It compares each scan to a cached snapshot and sends real-time update alerts via Gmail SMTP when new jobs are added or removed. The system also supports basic inbox polling via IMAP to receive user-issued commands such as "Apply to: [Job Title]", logging the intent to apply. Plan to add features for auto applying, tailoring resume and cover letter creation using AI according to job description.

Built as a mid-term project, for my for my CS634 Data Mining course, a Python-based tool that compares the Apriori algorithm with a brute-force method for mining frequent itemsets and generating association rules from simulated retail transaction data. It allows users to specify support and confidence thresholds, generates datasets of grocery item purchases and runs both algorithms to extract strong rules. Execution times are recorded to highlight the efficiency benefits of Apriori over exhaustive search. The project demonstrates practical applications of pattern mining for business intelligence.

Built as a final project, for my CS634 Data Mining course, a Python-based application that performs unsupervised clustering and outlier detection entirely from scratch. I generated synthetic 3D datasets, removed statistical outliers using a k-nearest neighbor distance method and then applied both K-Means and Hierarchical Agglomerative Clustering (HAC) with multiple linkage strategies. Each algorithm's results were evaluated using the Silhouette Coefficient to measure cluster quality. The program logs execution time, clustering effectiveness and dynamically determines the best method for each dataset.

Trip Sync is a real-time web application that visualizes global ship locations using live maritime data. It integrates Microsoft Maps and the Marine Traffic API to plot ships across interactive ocean maps. Both the frontend and backend are developed using JavaScript and PHP to handle API responses, dynamically render map markers and update ship positions based on user queries. This project focuses on real-time data synchronization, location tracking and responsive UI rendering.

Built a dynamic inventory management web app designed for real-time tracking and updating of items. The frontend was developed using React and Material-UI to create a responsive, component-based UI with modern styling. Next.js powered server-side rendering and routing, while Firebase handled real-time database updates and user authentication. I used both JavaScript and TypeScript for flexible and scalable client-side logic, ensuring seamless user interactions and efficient state management.

Verro is a fully functional e-commerce website targeting eco-conscious consumers that is built for the final group project for IT310. The website is created using wix and it features a clean responsive design, product search and filtering, secure checkout, a dynamic blog and a customer newsletter system. The platform supports real-time inventory management through Wix CMS collections and integrates seamlessly with Wix Payments for secure transactions.

Created a desktop-based admin dashboard to visualize monthly sales and key business metrics through interactive charts and reports. The application is built using Java and JavaFX for a responsive, user-friendly interface. Data is managed and queried using MySQL, enabling real-time insights and efficient decision-making based on sales trends and operational data.

Created an interactive web app using Dash and Plotly for data visualization and regression-based prediction for my CS301 Data Science class project. Integrated Hadoop for storing and preprocessing large datasets to enable scalable analysis. Users can upload CSV datasets, explore correlations and categorical trends through dynamic graphs, select features and train a linear regression model with real-time performance metrics. Enhanced results delivery by exporting processed outputs into Tableau dashboards for interactive and presentation-ready visualizations. The app includes built-in data preprocessing, one-hot encoding, imputation and scaling pipelines and supports on-the-fly predictions based on user input.