Staff Applied Scientist, Experimentation and Causal Inference

DoorDash, Inc.; California; Sur place; USD 180 000 - 240 000; Plein temps

Résumé du poste

DoorDash, Inc. is seeking a Staff Applied Scientist to provide scientific leadership for Curie, its Experimentation Platform within the Decision Systems team.

You will define and advance statistical, causal inference, ML, and AI capabilities powering high-quality decisions across DoorDash's three-sided marketplace. Your work will shape how teams design experiments, analyze results, interpret outcomes, and make faster, more trustworthy decisions.

Qualifications

• 6+ years of industry experience in statistical, causal inference, machine learning, or experimentation systems with measurable business impact.
• Deep expertise in experimentation and causal inference, with production experience in variance reduction, ratio metrics, sequential testing, heterogeneous treatment effects, marketplace interference, or quasi-experimental designs.
• Strong coding and data skills, including the ability to reason about production data pipelines and review complex analytical SQL for statistical correctness.

Responsabilités

• Set the scientific direction for Curie and shape the Decision Systems roadmap with engineering and product partners, identifying where better science, tooling, or automation can materially improve decision quality and velocity.
• Own the statistical contract for experiment analysis, including how metrics, populations, exposure windows, covariates, ratio metrics, and edge cases are defined.
• Ensure production systems faithfully implement those contracts by reviewing analytical SQL, validating statistical parity, and raising the bar for correctness across engine migrations.
• Develop and productionize advanced methods such as CUPED/CUPAC, ratio-metric and delta-method inference, sequential testing, heterogeneous treatment effects, switchback designs, and difference-in-differences for marketplace interference.
• Build approaches that hold up in high-scale, high-noise, three-sided marketplace data, including power analysis, multiple-testing correction, winsorization, quantile metrics, and other robust measurement techniques.
• Advance AI-assisted analysis workflows, including trustworthy automated experiment readouts, risk detection, and hypothesis generation, with the same rigor expected of statistical methods.
• Advise and mentor data scientists, product teams, operators, and engineers on experimental design, metric strategy, causal analysis, and interpretation of results.

Connaissances

Statistical & ML expertise

Causal inference

Experimentation systems

Production data pipelines

Outils

SQL

Python

Description du poste

About The Team

The Decision Systems team helps DoorDash teams make faster, more rigorous, and more trustworthy decisions. We build the analytical systems, statistical methods, and product experiences that power experimentation, causal inference, rollout analysis, metric interpretation, and business impact measurement across DoorDash’s three-sided marketplace. Curie is DoorDash’s experiment analysis product within Decision Systems. It sits at the center of how teams move from hypothesis to analysis to confident action, helping decision makers understand experiments, rollouts, metric movements, and marketplace impact while preserving scientific rigor, velocity, trust, and interpretability. Our work spans experimentation platforms, causal inference, variance reduction, switchback experimentation, metric systems, and AI-assisted analysis workflows. We partner closely with data scientists, engineers, product managers, and operators across DoorDash to translate advanced methods into reliable, scalable systems that improve decision quality and product velocity. Some of the work done in the team was published in articles such as: Design-Aware Variance Reduction for Switchback Experiments: A Comparative Study, Power-Optimal Covariate Adjustment for Switchback Experiments, CUPED on Steroids: Multivariate Covariate Adjustment for Switchback Experiments.

The Decision Systems team helps DoorDash teams make faster, more rigorous, and more trustworthy decisions. We build the analytical systems, statistical methods, and product experiences that power experimentation, causal inference, rollout analysis, metric interpretation, and business impact measurement across DoorDash’s three-sided marketplace. Curie is DoorDash’s experiment analysis product within Decision Systems. It sits at the center of how teams move from hypothesis to analysis to confident action, helping decision makers understand experiments, rollouts, metric movements, and marketplace impact while preserving scientific rigor, velocity, trust, and interpretability. Our work spans experimentation platforms, causal inference, variance reduction, switchback experimentation, metric systems, and AI-assisted analysis workflows. We partner closely with data scientists, engineers, product managers, and operators across DoorDash to translate advanced methods into reliable, scalable systems that improve decision quality and product velocity. Some of the work done in the team was published in articles such as: Design-Aware Variance Reduction for Switchback Experiments: A Comparative Study, Power-Optimal Covariate Adjustment for Switchback Experiments, CUPED on Steroids: Multivariate Covariate Adjustment for Switchback Experiments.

About The Role

We are looking for a Staff Applied Scientist to provide scientific leadership for DoorDash’s Experimentation Platform. In this role, you will define and advance the statistical, causal inference, machine learning, and AI capabilities that power high-quality decision-making across DoorDash. You will work on challenging problems at the intersection of experimentation, marketplace dynamics, machine learning, and product engineering.

Your work will shape how DoorDash teams design experiments, analyze results, interpret noisy marketplace outcomes, and make faster, more trustworthy decisions. This is a high-leverage role for a scientist who enjoys both deep technical work and broad cross-functional influence—partnering with engineers, data scientists, product managers, and operators to translate cutting-edge methods into reliable platform capabilities used across the company.

You're Excited About This Opportunity Because You Will…

• Set the scientific direction for Curie and shape the Decision Systems roadmap with engineering and product partners, identifying where better science, tooling, or automation can materially improve decision quality and velocity.
• Own the statistical contract for experiment analysis, including how metrics, populations, exposure windows, covariates, ratio metrics, and edge cases are defined.
• Ensure production systems faithfully implement those contracts by reviewing analytical SQL, validating statistical parity, and raising the bar for correctness across engine migrations.
• Develop and productionize advanced methods such as CUPED/CUPAC, ratio-metric and delta-method inference, sequential testing, heterogeneous treatment effects, switchback designs, and difference-in-differences for marketplace interference.
• Build approaches that hold up in high-scale, high-noise, three-sided marketplace data, including power analysis, multiple-testing correction, winsorization, quantile metrics, and other robust measurement techniques.
• Advance AI-assisted analysis workflows, including trustworthy automated experiment readouts, risk detection, and hypothesis generation, with the same rigor expected of statistical methods.
• Advise and mentor data scientists, product teams, operators, and engineers on experimental design, metric strategy, causal analysis, and interpretation of results.

We're Excited About You Because You Have…

• 6+ years of industry experience, or equivalent applied research experience, developing statistical, causal inference, machine learning, or experimentation systems with measurable business impact.
• Deep expertise in experimentation and causal inference, with production experience in methods such as variance reduction, ratio metrics, sequential testing, heterogeneous treatment effects, marketplace interference, or quasi-experimental designs.
• Strong coding and data skills, including the ability to reason about production data pipelines and review complex analytical SQL for statistical correctness.
• Experience translating statistical or machine learning methods into reliable, scalable platform capabilities in close partnership with engineering and product teams