Ally Financial Careers

As a key member of the Fraud Analytics, Strategies and Trends (FAST) team, the Principal Data Scientist plays an integral role in managing and optimizing vendor-owned fraud detection models, as well as driving data-driven fraud analysis. Using advanced analytical techniques, the Principal Data Scientist will develop insights to support cross-functional collaboration to implement policy and procedural changes that strengthen fraud prevention efforts.

• Manage, monitor, and optimize external fraud detection models provided by third-party vendors
• Collaborate with vendors to understand model performance, adjust thresholds, and ensure accuracy in detecting fraudulent activities
• Create and maintain regular reports and dashboards to track model performance, detect trends, and identify key performance indicators
• Assess and validate vendor model updates or changes, ensuring alignment with business needs and fraud detection objectives
• Perform regular audits and reviews of model output, providing insights into effectiveness and recommending improvements
• Lead or contribute to the development of new vendor model initiatives or data-driven tools to enhance fraud detection abilities
• Leverage advanced statistical and machine learning techniques to analyze large, complex datasets and generate actionable insights
• Develop analysis to support potential policy and/or procedural changes to reduce fraud and improve the customer's experience, and provide recommendations based on findings
• Develop new techniques for fraud analysis, as well as recommendations for internal FAST policies and/or procedures to help drive efficiencies and repeatable quality results.
• Support the continuous enhancement of fraud detection systems and policies to proactively mitigate risks
• Develop strong relationships with other FAST team members to drive collaboration, shared insights across the team, as well as coaching and mentoring peers toward improvement.
• Develop strong relationships with key business partners both inside and outside Fraud Prevention, pursuing solution delivery and shared successes.
• Deploy findings of analysis through effective presentation skills and use of influence.
• Deploy thorough controls that include not only well-developed countermeasures, but also detailed change control and backup documentation to ensure controls are understood and sustained.

• 5+ years of experience using advanced data manipulation techniques to solve business problems required
• Demonstrated experience managing the lifecycle of predictive models, including risk assessment, performance validation, and ensuring compliance with governance and regulatory standards required
• Experience with vendor model highly preferred
• Data preparation programs such as SQL, SAS, and/or Python.
• Solid understanding statistical and machine learning programming tools such as Python and R as well as hands on experience with model development. Experience with DataRobot and DataWalk is a plus.
• Proficiency in various database environments, such as Snowflake, Hadoop, or Oracle and proven experience to extract data.
• Linear and logistic regression, time series analysis, decision trees, survival analysis, ensemble trees (gradient boosting, random forest), and other quantitative techniques.
• Data visualization skills through Power BI (Preferred) or Tableau.
• Ability to analyze complex data, deriving clear action-oriented recommendations for the business, and communicating insights clearly and concisely to individuals from diverse backgrounds.
• Comfort level with ambiguity and ability to manage multiple projects at the same time.
• Degree in Data Science, Analytics, Computer Science, Economics, Statistics, Mathematics, Engineering or similar field of study (Master's or PhD preferred).
• Preferred - Experience working in financial services industry, specifically in fraud prevention.