Davit Arsenyan

Developed a machine learning system to predict which banking customers were most likely to take a loan within the next 30 days using behavioral and financial data. Built a feature engineering pipeline, achieved a ROC-AUC of 0.91 with LightGBM, and integrated SHAP explainability into a Flask dashboard for business users.

Developed a machine learning system to forecast air pollution levels in Yerevan using historical environmental data and real-time monitoring. The project includes data preprocessing, predictive modeling, and an interactive web dashboard that provides accessible air quality forecasts and visualizations.