Applied Scientist, Payments & Fraud Prevention, AWS Payments & Fraud Prevention

About the role

Are you passionate about building machine learning systems that protect one the world's largest cloud platform? The AWS Payments & Fraud Prevention (P&FP) Science team is looking for a driven Applied Scientist to help safeguard AWS and its millions of customers from evolving fraud threats. In this role, you will design, build, and deploy end-to-end machine learning models that detect and prevent fraudulent activity. You will work with massive, real-world datasets across the AWS payments, signup and usage ecosystem, develop new detection strategies, and take models from concept to production. You will also apply Generative AI (GenAI) techniques to enhance fraud signal discovery and strengthen our detection capabilities. At AWS, we process billions of transactions every day for hundreds of thousands of businesses worldwide. Fraud patterns shift constantly and our defenses must stay ahead. If you enjoy owning problems end-to-end, shipping models that make real-time decisions at scale, and making a direct impact on customer trust and financial protection, we invite you to join us and help shape the future of fraud prevention at AWS.

Responsibilities

• Design, build, and deploy end-to-end machine learning models and rules to detect, prevent, and mitigate fraudulent activities across the AWS payment and usage ecosystem.
• Source, extract, and analyze large-scale behavioral, transactional, and historical datasets to uncover fraud patterns and emerging threats.
• Apply hands-on expertise in statistical modeling, traditional machine learning, and analytics to identify and isolate issues across the fraud landscape.
• Explore and apply GenAI techniques, including large language models (LLMs) and synthetic data generation, to enhance fraud detection capabilities.
• Own the full model lifecycle — from data extraction and feature engineering through evaluation, productionalization, and deployment.
• Continuously monitor model and rule performance and improve robustness against adversarial behaviors and evolving fraud tactics.
• Experiment, prototype, and iterate on new detection strategies, algorithms, and evaluation metrics with a focus on rapid time-to-production.
• Collaborate closely with engineering, product, and operations teams to translate business needs into scalable