Principal Decision Scientist

Position Overview:

As a Decision Scientist within the Data & Analytics organization, you will collaborate with cross-functional teams to measure agentic impact, influence leaders with in-depth analysis, and strategically partner with Product Managers to shape the future of our work. You will work in a dynamic organization that sits at the intersection of analytics, AI, and automation, and is building the framework for the agentic enterprise in real time.

Core Responsibilities

• Agentic Evaluation Frameworks: Develop and scale methodologies to evaluate the performance, reasoning, and reliability of agentic workflows and AI systems.
• Causal Inference & Attribution: Build sophisticated models to solve complex attribution problems. You will distinguish between incremental gains and organic trends using various data science and modeling techniques at your disposal.
• Experimental Design: Lead the design and analysis of complex experiments, moving beyond simple A/B testing into multivariate testing, switchback experiments, and quasi-experimental designs for cases where randomization is impossible.
• Statistical Leadership: Refine and author methodologies, frameworks, analytical packages, and mentor junior team members.
• Strategic Influence: Translate complex statistical findings into clear, actionable narratives for executive leadership, influencing the long-term product and business roadmap.

Technical Requirements

• Background: Master’s or PhD in a highly quantitative field such as Statistics, Mathematics, Economics, Computer Science, Operations Research, or similar field.
• Experience: 10+ years of experience in a quantitative role, with a proven track record of deploying causal models or experimental frameworks in a production environment.
• Programming: Adept proficiency in Python or R (specifically the PyData stack: Pandas, NumPy, SciPy, Statsmodels, Scikit-learn).
• Data Retrieval: Mastery of SQL for complex data extraction, feature engineering, and performance tuning within cloud data warehouses (e.g., Snowflake, BigQuery).
• Modeling Mastery: Deep experience in high-dimensional regression, time-series analysis, and forecasting techniques.
• Agentic Systems: Familiarity with LLM evaluation metrics (LLM-as-a-judge) and the unique statistical challenges posed by non-deterministic AI outputs.