👩‍💻 Projects

🌤️ Astra Link to heading

Astra is an AI-powered weather and air quality monitoring assistant designed for asthma patients. The project integrates a React frontend and Flask backend with real-time data from the National Weather Service and AirNow APIs, enhanced by Google Gemini AI for personalized health recommendations. Features include an asthma-friendly responsive UI, geolocation and search via OpenStreetMap, and both web and CLI interfaces. Focused on building a scalable, health-centric application that combines data analysis, AI-driven insights, and accessibility to support proactive health management.

Skills:
Python, ReactJS, Flask, RESTful API, Google Gemini API

❤️ Heart Disease Prediction Link to heading

Built a full ML pipeline to predict heart disease from clinical data, tested and compared six models (LightGBM, XGBoost, SVC, MLP, CNN, and TabTransformer), used SMOTE and cross-validation, and found that LightGBM was about 80% accurate.

Skills:
Python, Scikit-learn, TensorFlow, PyTorch, Pandas, NumPy, Matplotlib, SMOTE, MinMaxScaler, Jupyter Notebook

📄 Open-Source Contribution Link to heading

Co-authored pull request (#4648) with the Cypress team to correct an incorrect command in the official Cypress documentation, improving usability and reducing confusion for developers.

Skills:
Documentation, Open Source, Cypress, Technical Writing

🕵️ Exploring Career Preparedness and Marketable Skills Among University Graduates Link to heading

This research project explores the career readiness and perception of marketable skills among graduating students at East Texas A&M University, aligning with the university’s strategic planning goals for 2024–2029. Utilized Python and Pandas to process and analyze Graduate Exit Survey data (Fall 2024 & Spring 2025), performing data cleaning, transformation, and statistical analysis. Computed mean and weighted mean self-assessment scores for each skill using Pandas, allowing for an objective assessment of student confidence in key skills. Visualized trends in skill rankings and confidence levels through Matplotlib and Seaborn, producing bar charts and heatmaps to illustrate the frequency and distribution of skill ratings. Applied descriptive statistics to explore the relationship between skill ratings and demographic variables, such as education level and gender, to identify key areas for curriculum enhancement. Performed the Mann-Whitney U test (non-parametric) to compare the distribution of self-assessment scores between independent groups, assessing whether there were statistically significant differences in skill perceptions between these groups.

Skills:
Python, Pandas, NumPy, scipy.stats, Matplotlib, Seaborn, Mann-Whitney U Test, Descriptive Statistics, Data Visualization, Data Cleaning

⏱️ Vibe Timer Link to heading

A full-featured, custom timer application for iOS built with SwiftUI. The default iOS timer continues to ring until manually stopped, which can be inconvenient when hands are occupied or users want a short alert. This app introduces a customizable timer that allows users to define how long the alert should sound before stopping automatically. Features include custom timer creation with an intuitive wheel picker, multiple simultaneous timers, customizable alarm durations and ringtones, timer presets for common tasks like “Egg Boil” or “Pomodoro”, background functionality with lock screen notifications, Live Activities displayed on the lock screen and Dynamic Island showing real-time countdown, and interactive controls to pause, resume, restart, or delete timers.

Skills:
Swift, SwiftUI, iOS, User Notifications, Live Activities, Widget Extension, Xcode

🔨 Skill Portal Link to heading

Skill Portal is a Resume/CV Selection and Evaluation System developed in Python to streamline the hiring process. It was built as a fulfillment of major project during undergraduate studies. The system ranks resumes and matches them with job descriptions using Cosine Similarity and NLP, providing a platform for uploading CVs and showcasing skills. Skills, education, and experience from CVs are vectorized based on job descriptions and compared using cosine similarity to rank candidates effectively. The platform enhances hiring efficiency and candidate evaluation, helping unemployed individuals present qualifications to potential employers.

Skills:
Python, Django, SQLite, NLP, Cosine Similarity, Data Processing, Resume Parsing

📚 Hamro Pustakalaya Link to heading

Hamro Pustakalaya is a web-based virtual library management system built in Java as a minor project during undergraduate studies. Features include online book search, availability checks, and reservation system, along with admin features like book inventory, user management, and fine tracking. The system includes a content-based recommendation system and user login and registration functionality.

Skills:
Java, T-SQL, mongoDB, JDBC, Spring Boot, Library Management System, Recommendation System

🧘‍♀️ +ve or -ve? Link to heading

+ve or -ve? is a desktop sentiment analysis application built with Java Swing and basic NLP techniques as a project to showcase in LITE Technical Exhibition 2019 hosted by Kantipur Engineering College. The application provides real-time analysis of sentences for positive or negative sentiment, featuring an intuitive Java Swing UI with extensibility for future enhancements. Designed as a standalone application for feedback analysis.

Skills:
Java, Swing, NLP, Sentiment Analysis, Text Classification