Pooyan Amir Ahmadi

About Pooyan Amir Ahmadi

I build and own org-scale decision and measurement systems for marketplaces, platform…

Experience

1. Principal Economist

   Amazon

   Sep 2021 — Present

   Lead global econometrics and applied-science initiatives informing multi-billion-dollar pricing, forecasting, and network-expansion decisions across Amazon’s fulfillment and marketplace systems, driving improvements in customer and seller experience worldwide. Architect production-grade Bayesian, causal-ML, and reinforcement-learning frameworks using Python, AWS SageMaker, EC2, S3, and PyTorch infrastructure, powering strategic planning, real-time decision intelligence, and platform optimization. Partner with senior business, finance, and product leaders to translate complex uncertainty into scalable, explainable AI systems that drive measurable growth and efficiency across regions and networks— Strategic & Business Impact —Built enterprise-scale econometric and AI systems, including the Flywheel Simulator, integrating Bayesian state-space, causal-ML, and reinforcement-learning frameworks to inform multi-billion-dollar pricing, forecasting, and network-expansion decisions across global fulfillment and marketplace operations— Technical Architecture & Deployment —Designed and deployed production-grade forecasting and decision-intelligence pipelines using Python, AWS SageMaker, EC2, S3, and PyTorch, integrating econometrics, machine learning, and simulation methods to achieve explainable, real-time optimization and elasticity insights at global scale— Organizational Leadership & Culture —Led and mentored cross-disciplinary teams of economists and scientists, embedding causal-inference and Bayesian best practices across orgs and partnering with senior product, finance, and ops leaders to drive data-informed strategic planning and operational excellence.• Awarded the 2023 Quantitative Economics Best Paper Prize (2021–2022 cycle) by the Econometric Society for research in Bayesian and structural econometrics, with recognition featured in Amazon Science.