Principal Data Scientist, Quantitative Intelligence (Consumer Transaction Data)

Principal Data Scientist, Quantitative Intelligence (Consumer Transaction Data) Reports to: Head of Data Science About the role Our client, a leader in consumer purchase data and analytics, turns large-scale card transaction data into intelligence for brands, investors and agencies. They’re hiring a Principal Data Scientist to be the hands-on technical lead for the team that turns this data into statistical methods and machine learning models. This is an individual-contributor tech lead role. You’ll guide the work of several data scientists, with a path to people management. A partner team handles data cleaning, merchant resolution and categorization. This role picks up from there: making the data statistically sound and predictive, then shipping it. What you’ll do Own models end to end: design the methodology, build it, deploy it to production and monitor it. Build the statistical layer that makes analysis on the data trustworthy: panel weighting, balancing, normalization, cohorting and bias correction. Build predictive models from past behavior, such as spend prediction, category and brand trend forecasting, propensity and churn, and behavioral embeddings. Apply causal methods (incrementality, synthetic control, uplift) where the problem calls for them. Set technical direction, review the team’s work, and explain and defend methodology to stakeholders and clients. Must-haves Hands-on experience with credit card or card transaction data. A track record of shipping ML and statistical models to production yourself, not just analysis or prototypes. Strong production engineering in Python and SQL on a cloud data warehouse. A deep statistics foundation: sampling, weighting and calibration, bias correction, and judging when a result can be trusted. Predictive modeling and forecasting built on raw, transaction-level behavioral data. [8+] years of experience, including leading the technical work of other data scientists. Clear, concise communication of methodology. Nice to have Causal inference and experimentation, as a complement to ML. Representation learning and embeddings. Privacy-preserving ML and data clean rooms. This role is not focused on fraud, AML, credit risk or loss/reserve forecasting, and it is not a pure experimentation or A/B testing role. Keywords: ML Scientist, Statistical Modeling, Predictive Modeling, Paneling, Normalization, Cohorting, Causal Inference, Synthetic Control, Uplift Modeling, Transaction Data, Behavioral Modeling, Forecasting, Representation Learning, Privacy Preserving ML, Cleanrooms, Production ML Pipelines