Principal Associate, Data Scientist - Bank Customer Protection Debit & Claims

Capital One•McLean, VA

13h

About The Position

Data is at the center of everything we do. As a startup, we disrupted the credit card industry by individually personalizing every credit card offer using statistical modeling and the relational database, cutting edge technology in 1988! Fast-forward a few years, and this little innovation and our passion for data has skyrocketed us to a Fortune 200 company and a leader in the world of data-driven decision-making. As a Data Scientist at Capital One, you’ll be part of a team that’s leading the next wave of disruption at a whole new scale, using the latest in computing and machine learning technologies and operating across billions of customer records to unlock the big opportunities that help everyday people save money, time and agony in their financial lives. The Bank Customer Protection Debit & Claims Data Science team builds the machine learning models that help our customers spend safely and get back on track if an issue does occur with their payments. We are constantly looking for ways to get ahead of fraudulent actors and scams before they have a negative impact on customers by analyzing historical transaction activity, account usage, merchant patterns and other data for signals that something is amiss. We use a variety of techniques, including representation learning and gradient boosting machines, to build purpose-built models that power our real-time decision systems and adapt quickly to emerging attack patterns. This role will bring these methodologies to bear on the debit authorization fraud side of our team - stopping debit fraud in real time as each transaction is authorized - spanning the full modeling spectrum, from proven techniques like gradient boosting to the frontier-AI approaches, such as graph and sequence learning, that are shaping the next generation of fraud detection.

Requirements

Currently has, or is in the process of obtaining one of the following with an expectation that the required degree will be obtained on or before the scheduled start date: A Bachelor's Degree in a quantitative field (Statistics, Economics, Operations Research, Analytics, Mathematics, Computer Science, or a related quantitative field) plus 5 years of experience performing data analytics
A Master's Degree in a quantitative field (Statistics, Economics, Operations Research, Analytics, Mathematics, Computer Science, or a related quantitative field) or an MBA with a quantitative concentration plus 3 years of experience performing data analytics
A PhD in a quantitative field (Statistics, Economics, Operations Research, Analytics, Mathematics, Computer Science, or a related quantitative field)

Nice To Haves

Master’s Degree in “STEM” field (Science, Technology, Engineering, or Mathematics) plus 3 years of experience in data analytics, or PhD in “STEM” field (Science, Technology, Engineering, or Mathematics)
At least 3 years’ experience with machine learning for predictive tasks, especially classification on large, highly imbalanced datasets (fraud, risk, or anomaly detection)
At least 3 years' experience in Python and SQL.
Preferred: production-quality, tested Python (pytest, mypy, linting, CI/pre-commit) and experience processing large-scale data with Spark (Polars, Snowflake/Snowpark)
Experience building and tuning gradient boosting models (XGBoost, LightGBM, or H2O) and deploying them into real-time or production decision systems
Experience building automated modeling pipelines – orchestrating training, evaluation, and deployment as reproducible workflows with Kubeflow Pipelines (KFP) on Kubernetes (or comparable pipeline/MLOps tooling)
Experience with model backtesting, validation, and performance measurement - precision/recall and capture rates at low decline/alert volumes
Experience coordinating data science projects in cross-functional teams

Responsibilities

Partner with a cross-functional team of data scientists, analysts, software engineers, and product managers to deliver a product that measurably keeps our customers safe from fraudulent activities.
Leverage a broad stack of technologies — Python, Conda, AWS, H2O, Spark, SQL and more — to reveal the insights hidden within huge volumes of numeric and textual data
Build machine learning models through all phases of development, from design through training, evaluation, validation, and implementation
Flex your interpersonal skills to translate the complexity of your work into tangible business goals