KDnuggets : News : 2006 : n02 : item14 < PREVIOUS | NEXT >


From: Steve Harbick
Date: 11 Jan 2006
Subject: Redmond, WA: Fraud Metrics Analyst, Job Code: 147633 at Microsoft

Are you an expert in data analysis and want to get into an internet ecommerce business? Are you passionate about your work and believe you can make a difference?

If you are experienced in database analytics and systems, this may be the opportunity for you to get involved in one of the fastest growing areas in Microsoft. The Fraud Systems & Prevention team within Global Commerce Operations (GCO) is looking for a customer-focused, detail-oriented, results-driven business professional with great communication skills and a strong knowledge of databases and statistical techniques who also has a background in the payments and fraud prevention space. The GCO Group manages the operations of the ecommerce platform for Microsoft online businesses such as MSN, Xbox Live, bCentral, Office, and many others. The Fraud Systems & Prevention team in particular, is responsible for assuring that we build a solid foundation for our platform to support our global ecommerce operations. We are passionate about enabling risk free transactions and we are looking to complement our high caliber team with this specialized skill. The Metrics Analyst analyzes, optimizes, and implements computational rules used to monitor and control activity on Microsoft's ecommerce platform in accordance with internal risk and governance policies. Because of the number, complexity, and interdependency of rules, the optimization of rule sets is a non-trivial statistical, machine learning and text analytics problem that spans numerous databases. It takes a special candidate to find the ideal solution. If you believe you are that individual, contact us.


  1. Use statistical techniques to design predictive models that detect fraud, etc. Iteratively test, refine, and improve the models. In this role, the Analyst will be required to be expert in building predictive, statistically valid models.
  2. Implements and updates computational rules to monitor and enforce fraud prevention policies. Manage the day-to-day aspects of systems (SAS, data sources, workflow systems, etc.) to ensure that the modeling and production infrastructure is optimized. In this role, the Analyst will work closely with IT and other data/systems providers (internal and external) to enhance systems, trouble-shoot data issues, etc
  3. Improve rule performance by analyzing misses, hit rates, and customer feedback. Organize, document and simplify rules. Contributes to global rule set by building best-practice rules and understanding international needs and Payment types.
  4. Collaborates with relevant GCO team members to insure rules are in alignment with business needs
  5. Assists the Software Development and Program Management teams in testing and deploying rule set.


  1. Master's degree in Statistics, Operations Research, or Machine Learning (PhD a plus) and a minimum of 3-5 years' related work experience. Proven ability to use modeling, optimization, and machine learning or text classification algorithms.
  2. Successful execution of this role will require that the Analyst be an expert in database record manipulation (mainly SAS/SQL) and statistical modeling techniques (logit, decision-tree, neural networks, cluster analysis, etc. using SAS). The key measure of success will be the degree to which GCO is able to reduce fraud and other violations of MSN's user agreement as a result of better and more proactive data and monitoring.
  3. Additionally, the candidate must demonstrate high energy/creativity, a passion for analyzing highly complex data sets, strong communication and project management skills, strong debugging skills, an entrepreneurial spirit, a relentless customer-focus, a practical understanding of quantitative methods, and superb attention to detail.

To apply, please send resume in e-mail to: fspjobs@microsoft.com, with the subject: KDNuggets job application.

KDnuggets : News : 2006 : n02 : item14 < PREVIOUS | NEXT >

Copyright © 2006 KDnuggets.   Subscribe to KDnuggets News!