Amazon is seeking an experienced Data Scientist to drive scientific tooling for LTPF forecasts and plans. The role involves managing complex programs and collaborating with various teams to enhance customer engagement with forecasting outputs.
Responsibilities:
- You will develop causal inference models, automated explainability frameworks, and variance bridging methodologies that translate LTPF's forecasts and plans into actionable business intelligence
- Your work will enable leadership to understand why forecasts and actuals diverge, what is driving demand shifts, and how strategic decisions propagate through the planning ecosystem
- You will build automated Plan-vs-Actual and Actual-vs-Actual variance decomposition models that quantify the contribution of individual demand drivers to observed gaps across revenue, price, units, inventory, and capacity metrics at multiple granularities to serve audiences from working-level analysts to VP-level planning reviews cycles
- You will build and maintain a causal model library with standardized hypothesis generation and validation pipelines, applying techniques from causal inference, time-series econometrics, and Bayesian methods. Each model will include calibrated confidence scoring and reusable components that scale across worldwide marketplaces
- You will develop GenAI-powered narrative generation capabilities that synthesize quantitative variance outputs into human-readable performance summaries and design automated hypothesis ranking to determine which demand drivers are most responsible for observed forecast error