At NimbusStack, I trained and evaluated machine learning models to estimate elasticity and examine how changes in demand affect cost. I prepared business data, explored demand–cost trends, and created interactive visualizations to help the team examine results for pricing decisions.
I built an insurance claim payment predictor that benchmarked five regression models; XGBoost achieved an R² of 0.78. I deployed it as a Flask web application for real-time predictions.
I also developed a customer churn model with 93% accuracy and a fraud detection model with 99.97% accuracy. My projects use Python and Scikit-learn, alongside data preparation, feature engineering, and exploratory analysis.
