Data Scientist - User Risk Measurement - Remote

Overview

The Intelligence & Investigations (I2) team detects and disrupts abuse and strategic risk so people can use our products safely. Within I2, Strategic Intelligence & Analysis is building a first-of-its-kind user-risk measurement function: policy-grounded baselines, confidence intervals, and attribution that reveal how real users engage with frontier AI—and how that changes when we ship mitigations. We sit between Safety Systems, Data Science, Integrity, and Product, turning heterogeneous safety signals into decision intelligence that appears in executive dashboards, launch/post-launch readouts, and weekly briefs.

In Short

• Define the measurement framework for user-level risk across products and cohorts.
• Establish baselines and statistical confidence for core metrics.
• Build decision-ready reporting surfaces like executive dashboards and weekly briefs.
• Clean and organize ambiguous data from disparate sources.
• Create attribution and change-tracking to connect shifts in user behavior.
• Partner across teams to ensure coherent analytics entry points.
• Uphold quality, privacy, and governance in measurement processes.
• Monitor signals for emerging risks and anomalies.
• Communicate insights and trade-offs clearly to executives and teams.

Requirements

• 3–6+ years in data science, measurement/causal inference, or risk analytics.
• Strong in sampling, inference, uncertainty quantification, and probability theory.
• Proficient in Python and SQL; experienced with data warehouses.
• Ability to communicate complex estimators into clear actions.
• Nice to have: experience with Airflow, Databricks, survival analysis, etc.

Benefits

• Relocation support available.
• Opportunity to work in a cutting-edge AI research environment.
• Collaboration with cross-functional teams.
• Contributing to safety and integrity in AI deployment.
• Equal opportunity employer with a commitment to diversity.