I've built AI projects that address environmental data scarcity in Pakistan, including a generative AI research project that produced 10,000+ synthetic data points, expanded usable data by 300%, and improved air-quality prediction accuracy by 20%.
I also developed a CNN image classifier achieving 92% validation accuracy and a Random Forest churn model with a 0.86 AUC-ROC score, using Python, TensorFlow/Keras, Scikit-learn, Pandas, SMOTE, and Streamlit. Beyond technical work, I coordinated a 500+ participant campus event, improving engagement through cross-functional collaboration.
