Rohan Gaikwad

I’m a Data Scientist who has built end-to-end machine learning pipelines using Python, SQL, Scikit-learn, XGBoost, Power BI, and Tableau. I focus on feature engineering, predictive modeling, and model evaluation, with hands-on work in class-imbalance handling, cross-validation, and business-focused delivery.

During a 6-month training program with Spinnaker Analytics, I independently designed and delivered 6 end-to-end ML pipelines across classification, regression, and clustering on real-world business datasets. My classification work used ensemble methods (Random Forest, XGBoost) with SMOTETomek resampling to handle class imbalance, reaching up to 97.42% accuracy and 99.75% ROC-AUC. I also built specific solutions like Hotel Reservation Cancellation Prediction (93% accuracy, 95.58% ROC-AUC), Company Bankruptcy Prediction (99.75% ROC-AUC on an imbalanced dataset), and NYPD Shooting Incident Fatality Classification (85.08% accuracy, 90.69% ROC-AUC).

I round out predictive modeling with regression and analytics projects, including Salary Range Prediction with an XGBoost model (91.72% R²) and Solar Energy Production Prediction improved to 3.04 RMSE using region- and equipment-specific features with linear regression. I’m comfortable communicating insights through data storytelling and BI dashboards—building Power BI dashboards from raw CSV ingestion through cleaning/transformation (Power Query) to final visualization.