Director, Data Science

Numerator•Chicago, IL

About The Position

At Numerator, understanding consumers begins with one of the industry’s richest and most comprehensive views of real-world purchasing behavior. For more than a decade, Numerator has pioneered and led the science of representing consumer purchasing behavior across retailers, channels, brands, categories, demographics, and geographies. As Director of Data Science, you will lead the team responsible for Numerator's consumer panel — the quality, currency, and representativeness of the signal at the heart of our data. Your team owns how well the panel reflects the population, how we understand where it doesn't, and the methodologies that keep it accurate, current, and trustworthy. You'll partner closely with Market Science, Engineering, Product, and GTM to turn that signal into products and estimates clients rely on. We’re looking for a leader who enjoys challenging assumptions, translating client needs into innovation, and developing exceptional data scientists. You should be equally comfortable debating statistical methodology, mentoring a team, partnering with Engineering and Product, and explaining complex concepts in language that drives better business decisions.

Requirements

8+ years applying advanced statistical modeling to large, complex, real-world datasets
Experience leading data science teams, including developing managers or leading through other leaders
Deep expertise in statistical modeling, machine learning, and probabilistic reasoning
Strong grounding in panel, survey, or sampling methodology — weighting, representativeness, and measurement science
Experience working with noisy observational data and developing methods to improve data quality, representativeness, or inference
Experience developing production methodologies rather than exploratory analyses, and designing statistical systems rather than simply applying existing techniques
Strong programming skills in Python, with working knowledge of SQL
Excellent communication skills, with the ability to explain complex statistical concepts to diverse audiences, including external/client audiences

Nice To Haves

Experience with Bayesian statistics, hierarchical modeling, probabilistic inference, or state-space models
Experience with causal inference, experimental design, or behavioral modeling
Experience developing statistical methods for consumer, retail, marketing, or longitudinal datasets
A demonstrated ability to challenge existing methodologies and introduce new scientific approaches into production
A systems-thinking approach that considers how improvements in one area influence the broader measurement platform

Responsibilities

Lead, coach, and develop a team of data scientists and data science managers focused on consumer behavior, panel methodology, and measurement science
Grow the leaders on your team — developing managers, not only individual contributors — and raise the technical bar through mentorship, methodology reviews, and adoption of modern statistical approaches
Foster a culture of scientific curiosity, rigor, and peer review, where decisions are grounded in evidence and healthy debate rather than precedent or intuition
Build an AI-native data science team by thoughtfully integrating AI into research, experimentation, software development, and scientific workflows — helping scientists move faster while maintaining rigorous standards
Maintain regular engagement with clients to deepen understanding of their challenges, build confidence in Numerator's methodology, and keep client needs central to the team's priorities
Own the quality, currency, and representativeness of the panel signal end to end — advancing methods for attribution completeness, compliance, anomaly detection, and keeping panelist demographics current
Characterize how the panel represents — and mis-represents — the population, treating that as a first-class, proactive modeling discipline rather than reactive data cleanup
Steward the panel as a living asset, partnering across the business on recruitment, engagement, and panel health, and turning evidence of where the panel is thin into targeted, prioritized investment
Design methodologies that improve the stability, consistency, explainability, and scientific defensibility of Numerator's consumer data
Challenge existing approaches through first-principles thinking, and research, prototype, and productionize new statistical and machine learning methods where they strengthen the platform
Partner with Market Science to advance the statistical foundations of Numerator's products
Partner closely with Market Science, Engineering, Product, and GTM — contributing the deep panel understanding that powers the broader measurement system, and translating scientific advances into scalable product capabilities
Communicate complex technical concepts clearly to technical and non-technical audiences
Balance scientific rigor with practical business impact and client value