PROJECTS

Inventory Tracker

Tech stack:Next.js, React, Firebase
I built a fully functional inventory management system using Next.js and React for the frontend, with Firebase handling backend data storage and real-time updates. The application allows users to add, update, and delete inventory items, with changes reflected instantly thanks to Firebase’s real-time database. I implemented form validation, dynamic UI updates, and state management to ensure a smooth user experience. This project helped sharpen my skills in CRUD operations, React hooks, and Firebase integration, and gave me valuable insight into building scalable single-page applications.

AI Customer Support

Tech stack:JavaScript, OpenAI API
I developed an AI-powered customer support chatbot using OpenAI’s GPT models to handle user queries with natural, human-like responses. Initially, I encountered challenges integrating the API and managing request-response flows, especially handling rate limits, token usage, and edge cases like vague or incomplete prompts. Through deep research and debugging, I implemented robust error handling and optimized the request structure to improve performance and accuracy. This project strengthened my understanding of AI integrations, API lifecycle, and building intelligent, context-aware systems.

IBM Skillsbuild Application

Tech stack:Spring Boot, JSP, SQL, Spring Security | Group Project
This project was developed as part of a group collaboration, where we built a secure and dynamic web application to support users in managing their IBM SkillsBuild learning journey. Our goal was to create a platform that not only tracks course progress but also encourages user engagement through course reviews.
This collaborative web application helps users manage their IBM SkillsBuild learning journey. Built with secure authentication, it provides a structured dashboard to view and track course progress, as well as a space to share course reviews.
My Role & Contributions:

  • Designed and implemented the dashboard with categorized tables: Available, Started, and Completed Courses.
  • Added a real-time progress bar to visualize course completion beneath the tables.
  • Built the Course Review feature to allow users to leave and view feedback on completed courses.
  • Integrated Spring Security for login, registration, and access control.
  • Structured code using MVC with packages: controller, domain, repository, service, and config.

Team Collaboration:

  • Used Git for version control and collaborative development.
  • Held regular meetings and followed Agile development practices.
  • Worked together on task breakdown, feature integration, and testing.

This project strengthened my skills in secure web development and team collaboration, while deepening my knowledge of Spring Boot and backend architecture.

Recipe Intelligence: Data-Driven Recommendation & Taste Prediction

Tech stack:Python, BeautifulSoup, pandas, scikit-learn, matplotlib | Group Project
Developed as part of a team, this project involved building a full data analytics pipeline — from scraping real-world data to deploying machine learning models. We collaborated on designing systems for intelligent recipe recommendations and taste prediction using real cooking data from BBC Food.
My Contributions:

  • Implemented the web scraping engine using BeautifulSoup to extract structured data from dynamic recipe pages.
  • Led the exploratory data analysis and built visualizations to uncover insights on recipe popularity and rating distributions.
  • Built a cosine similarity-based recommender system using combined recipe features to deliver personalized suggestions.
  • Implemented a binary classifier to predict whether a recipe would be considered “tasty” based on its ratings.
  • Contributed to team code reviews, debugging, and aligning implementation with academic requirements.

Key Features of the Project:

  • Web Scraping: Automated collection of recipe data from BBC Food into a structured pandas DataFrame and CSV export.
  • Statistical Analysis: Performed bootstrapped confidence interval estimation and explored rating distributions.
  • Recommender Systems:
    • Used Cosine Similarity (Vector Space Model) for basic recommendations.
    • Implemented a second recommendation engine using K-Nearest Neighbors (KNN).
  • Evaluation: Compared both engines on coverage and personalisation using test users and qualitative analysis.
  • Classification: Created a predictive model to determine recipe tastiness using a binary label and tested model accuracy.