Sr. AI Risk Manager , Res-Q

Millions of Selling Partners trust Amazon's marketplace to grow their businesses, and hundreds of millions of customers depend on us every day. Behind that trust is a network of systems, tools, and workflows designed to detect and resolve fraud and abuse at scale — and our team builds them.

We are looking for a Sr. Risk Manager who builds risk assessment and insight frameworks powered by AI. In this role, you will study incoming escalations not to investigate individual cases, but to understand the risk landscape deeply enough to build the models, classification systems, and analytical tools that transform raw escalation signals into structured, actionable intelligence at scale. You will design and deploy AI-driven risk assessment pipelines that automatically categorize incoming escalations, surface emerging abuse patterns, and quantify systemic defects across hundreds of thousands of annual contacts. You will use Amazon's AI and ML infrastructure to build topic models, root cause classifiers, and risk scoring frameworks that give investigators, program teams, and senior leaders the insights they need to act faster and prevent abuse before it scales. You will study investigation workflows to determine where automation replaces manual effort, where model-driven augmentation improves human judgment, and where new data collection mechanisms are needed to close intelligence gaps. This role offers the opportunity to shape how Amazon protects seller and customer trust by building the AI-powered risk intelligence layer that sits underneath every escalation decision our organization makes. Key job responsibilities ● Design, build, and deploy AI-powered risk assessment and classification pipelines that automatically categorize incoming escalations against governed root cause taxonomies, enabling systematic defect identification and trend detection across the full escalation portfolio ● Mine risk signals from unstructured escalation data to surface emerging abuse patterns, systemic enforcement gaps, and high-impact defect clusters that inform prevention strategies and program priorities ● Support the development of machine learning models using Amazon Bedrock, SageMaker, and supporting infrastructure to automate root cause classification, risk scoring, and escalation triage ● Study investigation processes and escalation workflows to identify where AI-driven automation, augmentation, or new data collection mechanisms can replace manual effort, improve consistency, or generate new risk insights ● Develop analytical frameworks and visualization layers that translate model outputs into actionable risk intelligence for Risk Managers, program teams, and senior leadership ● Define, instrument, and monitor model performance metrics alongside operational metrics to measure classification accuracy, automation coverage, defect reduction, and the business impact of AI-driven risk insights ● Partner with engineering, product, operations, and science teams to integrate model outputs into centralized tooling infrastructure, ensuring AI-powered risk assessments are surfaced at the right point in escalation workflows ● Conduct rigorous analysis on model performance, classification drift, and false-positive/false-negative trends to drive continuous improvement of risk assessment frameworks ● Prepare and present data-rich risk papers that track key risk indicators, escalation trends, and model-derived insights, and influence roadmap prioritization by quantifying where applied science delivers the highest operational return About the team The Res-Q team operates as the central command for abuse-related escalations, conducting investigations and driving systemic improvements across Amazon's global stores. Through our various intake channels, we handle the most complex and sensitive cases that require expert judgment and cross-functional coordination. Why Res-Q? We're building the next generation of risk management—one