I built and deployed an English toxic-comment classifier using NLTK preprocessing, TF-IDF, and scikit-learn, with cross-validation, hyperparameter tuning, and threshold analysis. It achieved 95.86% accuracy.
As a Python and AI Tutor at SQI College of ICT, I taught Python programming and introductory AI through hands-on exercises. I also built deployed Streamlit applications for customer churn prediction and laptop price prediction, and I'm developing a PyTorch transfer-learning pipeline for pneumonia classification.
