Data Scientist
Current• Developed a quarterly forecast model to predict sales of Tableau using a probabilistic time-series model GluonTS hosted on AWS Sagemaker, which improved error by 10% as compared to existing models• Created a model factory (generating 50+ unique models) using DataRobot to provide guidance on territory evaluation during the annual sales planning process, speeding up time to decision by 2x• Prototyped a discounting model to provide recommendations for optimal discount on a given deal based on account history and deal size. Additionally, the model predicts likelihood of a deal to be won and key features that impact deal success.