Data Scientist
Amazon
Data Scientist
Amazon
Sells products online and rents cloud computing
Own the detection models and the measurement science that find fraudulent and compromised AWS accounts in real time. Join a team of scientists and engineers whose analysis decides how fraud is detected and stopped, and how fast.
The AWS Payments and Fraud Prevention organization keeps AWS a safe place to build for millions of customers. Our team owns the science behind fraud and compromise detection on AWS. Account compromise is the case where a legitimate customer's credentials are stolen or exposed and an attacker uses the account to run unauthorized workloads. Fraud is the case where an account exists to consume resources with no intent to pay. Most accounts our models score belong to legitimate customers, so every decision has to stop fraud while leaving their work undisturbed. As a Data Scientist you will own detection models end to end, from the first look at an emerging pattern through feature design, training, shadow evaluation, production launch and the weekly performance review. You will define how we measure ourselves: precision by outcome, time to detect, runtime of unauthorized resources, and the cost a model prevented.
- Own models that detect and block fraudulent activity in real time, protecting customers and AWS.
- Turn a pattern spotted in a handful of accounts into a model that screens millions.
- Define the metrics your models and systems are measured by and discuss them in weekly and monthly business reviews.
- End-to-end ownership, from the first signal to the enforcement action and the cleanup after it.
Key job responsibilities
- Own detection models and rules end to end: frame the problem with partners, source and engineer features from large-scale activity, sign-in and usage data, train and validate, run shadow evaluations, launch, and monitor performance after launch.
- Investigate new fraud and compromise patterns as they emerge, quantify their scale and design the detection response, from a fast rule to a trained model, with the trade-off between speed and precision stated.
- Define and audit the metrics that measure our systems: outcome-based precision and recall, time to detect, runtime of unauthorized resources, prevented cost. Own the analysis behind weekly and monthly business reviews.
- Evaluate agentic and LLM-based systems: build labeled evaluation sets, measure error rates in both directions, and set the thresholds under which automation is allowed to act.
- Extend detection to new AWS services as they launch, including AI services where spend accrues in minutes.
- Write clear technical documents for scientists and for non-technical partners in operations, engineering and leadership, and present them in science reviews.
- Mentor junior scientists and analysts on method, and help interview for the team.
A day in the life
You begin with the performance review of the models you own: what they caught, what was later overturned, and what slipped past every layer. You pull the misses, find that three share a pattern you have not seen before, size it in the data, write a shadow-mode rule, and put a one-page analysis in front of the team with the precision you expect and the cost of waiting for a full model. In the team's planning session for the next quarter, that analysis becomes a proposal for a new model and you make the case for its place on the roadmap against the other layers of the detection stack. You also audit the cases your agentic system resolved on its own and adjust its evaluation set and thresholds. You use AI coding assistants daily, for SQL, Python and documentation, and you check their work. Your stakeholders are the fraud engineering teams that power your models, the teams that act on their decisions, the AWS service teams whose customers we protect, and the leadership that reads your numbers every week.
About the team
We are a team of scientists, analysts, software and data engineers and a program manager, with close partners across AWS Fraud Prevention worldwide. We own the science behind fraud and compromise detection on AWS, from signal and detection through decision, enforcement, cleanup and reinstatement. We follow a layered detection strategy with online and offline machine learning models and agentic investigation systems. Our models also run in the AWS European Sovereign Cloud, where they must work under EU data-residency constraints.
We invest in the growth of every team member through mentorship, careful code and document reviews, and projects sized to stretch you. We celebrate promotions, and new members ship their first project within their first quarter.
- Master's degree in a quantitative field such as statistics, mathematics, data science, business analytics, economics, finance, engineering, or computer science
- Experience with data scripting languages (e.g. SQL, Python, R etc.) or statistical/mathematical software (e.g. R, SAS, or Matlab)
- Experience with statistical modeling / machine learning
- Experience in designing experiments and statistical analysis of results
- This role requires you to be a national of an EU member state.
- Experience with Machine Learning and Large Language Model fundamentals, including architecture, training/inference lifecycles, and optimization of model execution, or experience designing or architecting (design patterns, reliability and scaling) of new and existing systems
- Experience with one of the following areas: machine learning technologies, Reinforcement Learning, Deep Learning, Computer Vision, Natural Language Processing (NLP) or related applications
- Experience in fraud, risk, security or another adversarial domain where the data shifts in response to your models.
- Experience deploying machine learning models in cloud infrastructure and monitoring them after launch, including retraining.
- Experience with AWS services such as Redshift, SageMaker, Fargate, Lambda, Bedrock or DynamoDB.
- Experience framing an open-ended business question as a quantitative problem and owning the analysis end to end.
Amazon is an equal opportunities employer. We believe passionately that employing a diverse workforce is central to our success. We make recruiting decisions based on your experience and skills. We value your passion to discover, invent, simplify and build. Protecting your privacy and the security of your data is a longstanding top priority for Amazon. Please consult our Privacy Notice (https://www.amazon.jobs/en/privacy_page) to know more about how we collect, use and transfer the personal data of our candidates.
m/w/d
Our inclusive culture empowers Amazonians to deliver the best results for our customers. If you have a disability and need a workplace accommodation or adjustment during the application and hiring process, including support for the interview or onboarding process, please visit https://amazon.jobs/content/en/how-we-hire/accommodations for more information. If the country/region you’re applying in isn’t listed, please contact your Recruiting Partner.
Amazon runs the largest online store in the United States and rents cloud computing to businesses and governments through Amazon Web Services. AWS pioneered renting computing power over the internet, and Amazon built its own delivery network to get orders to customers quickly. Shoppers buy from Amazon and from independent sellers on its store, and Prime members pay a subscription for fast free shipping, streaming video and music, and other perks. AWS rents out computing power, storage, databases, and AI tools so companies can run software without owning servers, and Amazon also makes Kindle e-readers and Echo speakers with the Alexa assistant. Jeff Bezos founded Amazon in 1994, and it is based in Seattle, Washington.
Explore more opportunities at Amazon