Data Scientist

Numerator’s Data Science team provides statistical and methodological leadership across the organization, developing the methodologies, tools, and data products that support our products.

This is a wide-ranging applied data science role focused primarily on the consumer purchase panel underlying our largest products, with opportunities to contribute to other complex data science initiatives across Numerator’s broader product portfolio. The work combines applied statistics, data investigation, methodology development, and practical implementation. Many of the questions we tackle do not have a single observable right answer, requiring us to combine statistical evidence, product context, and subject-matter judgment to understand tradeoffs and determine whether a solution improves the product as a whole. As your expertise grows, you will take on increasingly complex work and have opportunities to deepen your technical expertise, lead projects, improve team practices, and shape how we work.

• Design and implement statistical methodologies that improve the representativeness, stability, and quality of Numerator’s products.

• Develop data science solutions to problems involving consumer behavior, panel composition, and data quality using techniques such as sampling, weighting, statistical modeling, classification, predictive modeling, anomaly detection, and optimization.

• Build scalable Python and SQL solutions, internal tools, and data products that turn statistical methodologies into reliable production workflows.

• Develop metrics, monitoring, and validation frameworks to evaluate system performance and identify unexpected or unintended behavior.

• Contribute throughout the project lifecycle, from problem definition and exploration through implementation, testing, deployment, and ongoing evaluation.

• Work closely with Product, Engineering, Data Operations, and business partners to shape requirements and develop solutions that are both statistically sound and practical to use.

• Communicate methodologies, findings, assumptions, and tradeoffs clearly to technical and non-technical audiences.