Staff Software Engineer, Data - Experimentation

Role overview Provide scientific and technical leadership for an experimentation platform used to evaluate product changes and business decisions in a complex, three-sided marketplace. You’ll shape the platform’s methods and roadmap, ensuring teams can design sound experiments, interpret uncertain results, and act with confidence. The role combines deep statistical work with production engineering and broad advising across functions.

Responsibilities - Set scientific direction for experiment analysis capabilities and help prioritize the platform roadmap with product and engineering partners. - Define the statistical rules for metrics, populations, exposure windows, covariates, ratio measures, and edge cases. - Review analytical queries and validate that production systems implement statistical definitions correctly, including during engine migrations. - Develop and deploy methods such as variance reduction, sequential testing, heterogeneous treatment analysis, switchback experiments, and quasi-experimental approaches. - Improve measurement quality for noisy, large-scale marketplace data through power analysis, multiple-testing controls, and robust metric techniques. - Advance AI-assisted experiment analysis, including automated readouts, risk detection, and hypothesis generation. - Advise and mentor technical and operational partners on experimental design, metrics, causal analysis, and interpretation.

Requirements - At least six years of relevant industry experience or equivalent applied research, with measurable impact from statistical, causal inference, machine learning, or experimentation systems. - Deep expertise in experimentation and causal inference, including practical experience bringing advanced methods into production. - Strong coding and data skills to reason about production data systems and analytical implementation. - Ability to work across scientific, engineering, product, and operational teams and guide technical decisions.

Nice to have - Experience with marketplace interference, ratio metrics, variance reduction, sequential testing, or quasi-experimental designs. - Familiarity with trustworthy AI-assisted analysis workflows.