Hire Andrew K. - Data Scientist

* Built hierarchical ML models for sales forecasting. Reduced forecast error to <2% and clarified business guidance for finance and marketing partners, highlighting the impact of economic and marketing factors across countries, products, brands, and metrics * Productionized data processing, model retraining, and monthly sales forecasting in Airflow * Developed event adjustment process for legacy forecasting process so that business teams could compare day of week/month/year impact on year over year sales growth * Analyzed impact from non-static holidays to determine how booking patterns would change in the following year * Built Tableau dashboards for sales forecasting, model comparison, and error analysis, enabling leadership to evaluate and select models by business need at any country, product, brand, or metric level * Created and taught SQL courses for summer interns

* Produced machine learning model to predict commercial real estate valuations. Harvested * data from multiple sources to provide a more robust property analysis * Created deep learning model for doctors to classify atrial fibrillations from * electrocardiograms to quickly diagnose and treat patients * Built movie recommendation demo and compared machine learning and deep learning * modeling techniques in Spark to demonstrate hardware performance for technology hardware * manufacturer * Constructed predictive model for a bank’s new account fraud system using anomaly detection * and classification models. Improved business rules for check fraud detection resulting in * significant decrease in manual review time. Developed loan default classification model * using IBM Cloud Pak for Data * Designed demand forecasting model for a food service supplier to a large fast-food chain for * limited time offer promotional items. Produced clustering model to determine proxy * promotional items based on ingredients