Lazr Galstyan

Data Scientist @Warner Bros. Discovery

Los Angeles, CA, US

Developed machine learning models to forecast visitor numbers for WB Studio Tour Hollywood with~94% accuracy, helping the Tour team to optimize workforce allocation and cut costs by up to 10%• Leveraged association rules and word2vec models to identify complementary products at the TourStore, creating product bundles that increased gross income by up to 11%• Applied transformer models (SBERT) and LLMs (GPT-4o) to analyze movie scripts for box officeprediction, empowering the Motion Picture Group to make data-driven movie greenlight decisions• Built data pipelines in R-Shiny application to ingest, validate, and store TV production cost data on acentralized server, enabling real-time comparisons of actuals, budgets and EFCs across multiple shows

Duke University - The Fuqua School of Business

Master of Science - MS, Quantitative Management: Business Analytics