About this role
Amazon's Relay Product Excellence team is responsible for protecting the integrity of Amazon's freight transportation network. Our team operates at the intersection of risk detection, carrier trust, and operational enforcement, ensuring that millions of loads move safely and reliably across the network every day. We investigate flagged cases from advanced detection models spanning theft, unauthorized subcontracting, carrier identity mismatches, and insider threats. Our work directly reduces fraud, strengthens carrier compliance, and feeds critical intelligence back into the systems that protect Amazon's supply chain. As the network grows in complexity and scale, so does the sophistication of the risks we manage, and we are looking for sharp, detail-oriented investigators to join us. We are hiring Risk Analysts who will serve as frontline investigators across our detection model portfolio. In this role, you will investigate cases flagged by machine learning models, make enforcement decisions on carriers and loads, and provide structured feedback that drives continuous improvement in detection accuracy. You will work across multiple risk workstreams including theft, subcontracting, fleet identity, and anomaly detection, collaborating closely with Business Analysts, Product, and Tech teams to close the loop between investigation outcomes and model performance. Key job responsibilities Investigate flagged cases and make enforcement decisions including suspend, escalate, or close. Tag false positives with structured reason codes, providing frontline intelligence that enables model improvement. Communicate patterns observed in the case queue through feedback syncs, WBRs, and per-case tagging. Own day-to-day execution of investigation workflows, resolving process gaps in tools and SOPs. Collaborate with Business Analysts on case prioritization by validating operational feasibility of proposed changes. Support false negative analysis and escalation deep dives with case-level expertise. Maintain productivity standards including cases per day, SLA adherence, and backlog management. A day in the life You come in, pick up your assigned case queue, and start working through investigations. You review flagged cases, pull the relevant data, make a call on whether to enforce or close, and tag your reasoning. If something looks off with the model or the queue (wrong cases surfacing, patterns that don't make sense, tool bugs), you flag it through the appropriate channel so it gets addressed. You hit your daily case target, keep your backlog clean, and move on. It's heads-down investigative work with a direct feedback loop into the systems you rely on.