Data Scientist
CurrentBuilt a demand forecasting model to predict daily sales for 100 products at over 4000 stores nationwide, amounting to over 4.5 million predictions per day. The model has improved accuracy by 4%, helping reduce stockouts.Moved the demand forecasting model I built into production. This was the company's first data science model to be moved into production. This involved collaborating with the ops team to create the production environment and implement the technology needed for running the model.Developed the company's change management process for moving data science models into production. This included coding standards, unit testing, version control, code walkthroughs, and testing in the stage environment.Trained and deployed a deep learning model for natural language processing. The model was part of a pipeline I built to parse 4.6 million company documents, and identify which ones would be needed by a part of the business that was being spun off.Designed a product recommendation system for customers in the hospital, K12, and university segments using both internal and external datasets. The sales team has been using this system since 2021, and in the most recent year, it caused a 47% increase in incremental sales.Collaborated with the relevant business unit for each project I worked on, to understand their needs, and to get their feedback throughout the project, helping us stay aligned.Created and delivered a summary presentation to business leaders for each project I worked on, to show the results, highlight key takeaways, and discuss next steps.Coached the junior data scientist on my team, providing guidance about how to turn a business question into a data science project, how to use coding best practices to ensure reproducibility, and how to talk to non-technical audiences.
