I built a customer churn prediction pipeline using Python, Pandas, Scikit-learn, and XGBoost, working with more than 7,000 customer records. By addressing class imbalance with class-weighted logistic regression, I improved churn-class recall from 58% to 84%.
As a Web Developer Intern at NeuAI Labs, I built a responsive grocery e-commerce website with React.js and Firebase Firestore. I’ve also developed a JavaFX cafe management system and an admission enquiry chatbot using IBM Cloud.

