Sr Risk Manager, Inventory Trust

The Selling Partner Trust and Store Integrity (TSI) org is committed to our vision that bad actors cannot operate in our store while selling partners start and grow their business without fear of disruption, such that customers and selling partners across the globe trust us, and have confidence in the integrity of Amazon's store.

As part of TSI org, the Inventory Trust (IT) team is

responsible for detecting and preventing inventory from bad actors reaching our customers. We build proactive detection mechanisms, risk models, and intelligence-driven processes to identify bad actors before they impact customers.

We are looking for a Risk Manager who thrives at the intersection of data-driven detection and investigative deep dives. In this role, you will lead efforts to improve detection model accuracy, build rapid-response risk rulesets for emerging modus operandi (MOs), investigate bad actor networks end-to-end, and partner with science and engineering teams to close detection gaps. You will work across Supply Chain Risk Detection, ORC investigations, and external intelligence programs to ensure our mechanisms stay ahead of evolving bad actor tactics.

The ideal candidate is a strategic thinker with deep investigative experience, capable of building and scaling sophisticated detection programs to combat bad actor networks.

Key job responsibilities

Lead deep dives on detection model performance (precision, recall) to identify signal gaps, false positive/negative drivers, and improvement opportunities for existing risk models.

Develop recommendations for feature engineering, threshold tuning, and signal integration.

Own end-to-end root cause analysis on detection defects, false positives, reinstatements, and escalations to identify systemic gaps in detection logic

Analyze escalation patterns to identify detection blind spots and prioritize coverage expansion.

Build and deploy quick-turn risk detection rulesets and heuristics to address newly identified bad actor MOs, emerging fraud patterns, and intelligence leads before models can be retrained. Translate external intelligence (ORC referrals, brand leads, cargo theft data, stolen serial number matches, law enforcement tips) into actionable detection queries and monitoring rules.

Continuously monitor and refine detection rulesets based on hit rates, enforcement outcomes, and evolving bad actor tactics; sunset rules that lose effectiveness.

Perform full-sweep investigations on bad actor rings by casting wide nets from early leads, expanding from a single enforcement to identify connected sellers, shared suppliers, common locations, financial linkages, and secondary user patterns.