Yasaman Shahi

Experience

1. Principal Data Scientist

   Walmart

   Mar 2024 — Present · Sunnyvale, CA, US

   Led the end-to-end redevelopment and deployment of Walmart’s VizPick Backroom Expected Picks Forecasting system, now live across all U.S. stores. Drove the full lifecycle—including data processing, model design, validation, production deployment, weekly forecast execution, and operational support. Collaborated closely with product and business teams on stakeholder alignment and requirements scoping. Improved forecast accuracy by 51.6%, enabling an estimated $206M in annual labor cost savings through optimized labor planning- Designed and deployed Walmart’s short-notice absenteeism forecasting tool (“Call-in”), predicting associate absences at the store-day, labor-role, and workgroup levels. Leveraged over 350 engineered features to improve forecast accuracy by 8.7%, enabling proactive staffing decisions and minimizing operational disruptions- Designed and led the Cohort Analysis initiative; applied causal inference methods to uncover root causes of labor under- and over-utilization across stores. Identified key drivers such as scheduling inefficiencies and operational breakdowns, enabling targeted, data-driven interventions that improved workforce productivity and store performance- Initiated and currently leading the development of the Agentic AI framework to detect patterns of potential misuse in short-notice call-in behavior. The system evaluates historical attendance data in context—peer norms, tenure, and scheduling trends—to flag elevated risk of excessive absenteeism or policy violations.