Description

Are you passionate about applying machine learning, time series forecasting, and operations research to transform the delivery of heavy and bulky items for Amazon customers? Are you excited about working with large-scale operational data and developing models that drive real business impact? If so, the Amazon Extra Large (AMXL) Science team may be the right fit for you.
AMXL is Amazon's specialized business for delivering heavy and bulky items — appliances, furniture, fitness equipment, and mattresses — with a premium customer experience that includes room-of-choice delivery, at-home installations, and assembly services. In this role, you will leverage large-scale operational data to develop and deploy predictive models and optimization solutions that solve real-world logistics and fulfillment challenges, partnering closely with scientists, engineers, and business stakeholders.


Key job responsibilities

Apply machine learning, statistical modeling, time series analysis, and operations research techniques to build solutions for delivery routing, capacity planning, demand forecasting, workforce scheduling, and network optimization

Analyze large-scale historical and real-time operational data to surface efficiency patterns, bottlenecks, and emerging trends across the AMXL network

Develop, validate, and deploy models that improve cost-to-serve and customer experience

Partner with cross-functional teams to implement data-driven strategies and measure impact

Build scalable, automated pipelines for data ingestion, feature engineering, model training, and validation

Monitor deployed model performance and communicate results through clear reporting on key operational and business metrics


About the team

You'll be part of a small, collaborative team of scientists who move fast and care deeply about the problems they solve. A typical week might involve whiteboarding a new forecasting approach with a senior scientist, partnering with engineers to push a model into production, deep-diving into operational data to understand why a metric moved, or presenting your findings to business leaders who will act on them. The work is high-visibility and high-impact. The models you build will directly influence how millions of heavy and bulky items reach customers.