Projects I worked on

AI Learning Platform (Final Year Project):

"I developed an AI-powered adaptive learning system that generates quizzes dynamically and tailors assessments to each student's performance. The platform was built with FastAPI on the backend, Next.js on the front-end, and PostgreSQL for data storage, using async SQLAlchemy for responsive database access and JWT authentication for secure user sessions. The result is a modular platform for personalized coding education that is ready for scalable, API-driven deployment."

Live Demo

Social Media Backend API:

"I engineered a complete social media backend using FastAPI, providing full CRUD functionality and JWT-based authentication for secure user access control. The PostgreSQL database was architected with the SQLAlchemy ORM and versioned through Alembic migrations, including a post-voting system built on composite primary keys. I also set up a CI/CD pipeline with GitHub Actions and Docker that automates integration testing through Pytest and handles production deployment."

Movie Recommendation System:

"I built a movie recommendation system that suggests titles to users based on their viewing preferences. Working in Python with Pandas and Scikit-learn, I cleaned and explored the dataset, engineered features from movie attributes, and applied similarity-based modelling to rank the most relevant recommendations for a given title. The finished model is deployed as an interactive Streamlit app, so anyone can pick a title and see recommendations instantly. This project deepened my understanding of recommender systems, feature engineering, and evaluating model quality on real-world data."

Live Demo

Student Score Prediction:

" The student score prediction project uses machine learning to forecast academic performance from factors such as study habits and prior results. I handled the full workflow in Python, from data cleaning and exploratory analysis with Pandas and Matplotlib through to training and evaluating regression models in Scikit-learn, comparing their accuracy to select the strongest performer. The trained model is deployed as a Streamlit app where users can enter their own inputs and get a predicted score in real time. It strengthened my skills in data analysis, regression modelling, and communicating insights through clear visualisations."

Live Demo