Sr. Data Scientist, Online Experimentation

About the role

The Home Depot is looking for a Senior Data Scientist in Online Experimentation. The role owns the statistical foundation of how the Online organization learns from its customers: the inference methods behind experiment results, the diagnostics that catch a broken test early, and the design guidance that leads product teams to well-powered, well-instrumented experiments. It is a senior individual contributor role with broad technical influence, reporting to a manager or above with no direct reports.

Key responsibilities

• Define and validate the statistical approach across online experimentation: estimator choice, variance reduction, sequential testing, heterogeneous treatment effects and causal inference where randomization is not possible.
• Work with product managers, engineers, designers and merchants to shape hypotheses, size effects honestly and interpret results without overclaiming.
• Design, build and evaluate LLM-based and agentic workflows that automate high-volume work such as design review, instrumentation checks, readout drafting and anomaly triage.
• Develop algorithms and models on large datasets, prepare reports and present insights to technical and non-technical leaders.
• Support project teams, mentor junior colleagues, help distribute workload and resources, and contribute to recruiting.
• Share education on advanced analytics with business partners and explore new ways to use data science as a competitive advantage.

Required profile

• Age 18 or older and legally permitted to work in the United States.
• Preferred: advanced degree (MS or PhD) in Statistics, Economics, Biostatistics, Computer Science, Applied Mathematics or a related quantitative field.
• Preferred: 6+ years of quantitative analytics experience, including hands-on ownership of online controlled experiments end to end.
• Preferred: experience mentoring analysts or data scientists and setting technical standards without formal authority.

Required skills

• Experimental design, hypothesis testing, power analysis and causal inference.
• Variance reduction, sequential and always-valid testing, and heterogeneous treatment effect estimation.
• Quasi-experimental methods such as difference-in-differences, synthetic control or geo-based design.
• Python or R for statistical analysis and SQL against a cloud warehouse (preferably Google BigQuery).
• LLM workflows: prompt design, tool and function calling, retrieval, multi-step or agentic orchestration, and evaluation of AI systems.
• Modern experimentation platforms, including server-side testing and feature flags.
• Building or validating an experimentation statistics engine, and familiarity with e-commerce metrics such as conversion, AOV, attach rate and LTV.