Amazon’s Last Mile is one of the world’s most complex logistics engines and one of the company’s fastest-evolving innovation frontiers. Every package delivered to a customer is powered by thousands of decisions made upstream in planning, routing, staffing, and execution. Our team’s mission is to transform these decisions using science, causal inference, and AI-driven reasoning. We are looking for a Senior Data Scientist who is passionate about building causal inference models, designing large-scale descriptive analytics frameworks, and advancing the next generation of GenAI-powered decision systems. You will play a pivotal role in shaping the science layer behind Amazon’s Last Mile Knowledge Graph, agentic AI assistants and analytics platforms that support operators and leaders across the globe. This role is ideal for a scientist who wants to go beyond models someone who wants to influence product strategy, build durable scientific foundations, and deliver measurable business impact at billion-dollar scale. You will experiment, innovate autonomously, and collaborate deeply with science, engineering, and product teams to reinvent how analytics and decisions are made. If you are excited by ambiguous problems, motivated by business impact, and energized by building science systems that shape operational decisions for tens of thousands of Amazon employees every day—this is the role for you.

Lead causal inference and descriptive analytics science for Last Mile, designing models that explain why things happen, not just what happened.

Develop experimentation frameworks, measurement strategies, and model evaluation pipelines that integrate directly into LM-Pulse, LM-YTP, and the Last Mile Knowledge Repository (KR).

Design and productionize predictive and prescriptive models that guide decisions in demand planning, labor strategy, operational health, and performance optimization.

Build science components that power GenAI-based products, enabling intelligent reasoning, explainability, and autonomous agent behavior.

Partner with DEs, BIEs, PMs, and Ops leaders to build scalable data and science architectures that support ML models, causal estimators, ontology-driven insights, and KG integrations.

Drive experimentation at scale, identifying high-impact opportunities where modeling can directly reduce cost, improve speed, or enhance customer experience.

Influence product and science strategy by defining scientific tenets, challenging assumptions, and raising the bar on analytic rigor across the organization.

Communicate scientific findings clearly and persuasively to VP-level audiences, shaping decisions across planning, execution, and network strategy.