Amazon’s Customer eXperience Impressions (CXI) team is hiring a Software Engineer to develop machine learning-driven decision systems that detect defects in the Amazon shopping experience and recommend interventions. This role requires expertise across the full machine learning lifecycle, including feature engineering, model development, inference optimization, and real-time decisioning. The engineer will work on large-scale data pipelines, recommendation systems, and automated decision frameworks to ensure Amazon acts before friction impacts demand and customer visits.
CXI operates at the intersection of customer experience and supply chain optimization. As part of Supply Chain Optimization Technology (SCOT), the team’s data infrastructure captures signals from every stage of the shopping journey—search, detail page, and repeat purchase interactions—to measure when and why customers hesitate, abandon their carts, etc.
The engineer will build real-time ML pipelines that evaluate these signals, to ensure that interventions happen at the moment they can still recover lost sales. This includes developing and deploying large-scale ML models that detect and rank defects based on severity and intervention confidence, determining when and how Amazon should act.
Key job responsibilities
This engineer will work closely with Scientists to ask research questions about customer behavior, develop models for defect detection and intervention prioritization, and design experiments to validate assumptions. They will integrate machine learning models into high-performance, distributed decision systems, to ensure low-latency execution at Amazon’s scale. Through experimentation and iterative model refinement, they will ensure that interventions improve purchase outcomes while avoiding unnecessary corrections.
Beyond model integration, this role requires expertise in real-time inference optimization and distributed model serving. The engineer will optimize feature stores, build online learning mechanisms, and design architectures that adapt intervention strategies dynamically based on observed outcomes. They will also develop causal inference frameworks to quantify the true impact of delayed or missed interventions, to ensure the system learns from past decisions to improve future accuracy.
This engineer will actively participate in the Amazon ML community, sharing best practices and mentoring software development engineers who have an interest in ML. Their work will directly benefit customers and the business by ensuring Amazon not only detects shopping experience defects but also determines when and how to act with precision. Their contributions will drive real-time corrections, preserving demand and improving customer satisfaction at scale.
This is a highly visible role and will require regular interaction and communication with senior leaders. Excellent written and verbal communication skills is important.
