Data Scientist - Consumer Decision Sciences
American Express
We are looking for a data scientist for the Decision Sciences team within the American Express US Consumer Card Services Group.
Risk and Information Management (RIM) comprises several teams. Risk Management, the largest of the teams, manages the Company's Credit, Market, and Operational Risk. The team works with each business unit globally, building best-in-class capabilities and processes to manage and monitor risk prudently and profitably. The Risk Management team's work extends across the customer life cycle, including aiding with product design, identifying profitable prospective customers, defining new account approval criteria, determining effective and efficient cross-sell methods and point of sale practices, and setting strategies for collections.
Job Responsibilities:
You will join the Machine Learning Research and Digital Acquisition Modeling team and have responsibility for identifying, leveraging, and enhancing statistical learning algorithms relevant for a diverse set of problems, with a focus on understanding data arising from online consumer advertising and engagement. You will analyze large and sometimes messy data in order to develop insights into customer behavior and introduce new approaches to transform complex behavioral data into actionable information, such as building predictive models to improve our decisions.
Sample projects:
- Analyze large-scale sparse datasets to identify target populations and improve offer relevance.
- Hyperparameter optimization for machine learning algorithms such as k-nearest neighbor, gradient boosting, and ensemble methods.
- Prototype and simulate streaming test and learn algorithms.
Required Skills/Qualifications:
You have a passion for empirical research, a desire to work on challenging data problems around the digital advertising ecosystem, and a demonstrated ability to learn and innovate. You are comfortable communicating and receiving feedback about your work with both technical and non-technical audiences. Research in an industrial or academic setting. Deep understanding of machine learning and statistical modeling algorithms such as logistic regression, k-nearest neighbors, gradient boosting, random forests, or collaborative filtering.
- Ability to transform data and prototype quickly to conduct statistical analysis using tools like R, Python, Java, or SAS.
- Proficiency with SQL and relational databases.
- Proficiency with Unix/Linux environments.
- Familiarity with Hadoop environments and data tools such as Hive or Pig.
New York City
Full Time