Quantitative Intelligence Analyst

OpenAI•San Francisco, CA

4d•$198,000 - $320,000•Hybrid

About The Position

As a Quantitative Intelligence Analyst, you will focus on discovering novel and emerging risks in complex human–AI systems before they are well-defined, measurable, or widely understood. You will use deep subject matter expertise and quantitative tooling to surface weak, early, and unconventional risk signals. You will build analytic models that explain how harms could emerge and translate ambiguous patterns into structured, data-driven insight. Your work will help identify potential gaps in policy or coverage and operationalize previously unmeasured problems into signals that can support detection, mitigation, and planning downstream. You will develop analytical frameworks that map how new risks form, evolve, and propagate as products change, policies shift, and external events unfold. Your analyses will directly inform strategic risk prioritization and planning across the company, with regular visibility through strategic risk products. This role is based in office (hybrid, 3 days/week). Relocation support is available.

Requirements

3+ years of experience in quantitative intelligence analysis, trust & safety, security analysis, or risk-focused research
Comfortable working on complex trust and safety domains such as child safety, violent activities, self-harm, or similar high-stakes risk areas
Familiarity with data mining, statistical modeling, and supervised learning methods
Understand how to monitor signals or models for data drift, behavioral adaptation, or performance degradation over time, and can diagnose likely causes
Experience in operationalizing adversarial or strategic risk behaviors, including through red-team exercises, agent-based modeling, or structured scenario analyses
Comfortable working with Python and SQL

Nice To Haves

Experience with quantitative stress testing or Monte Carlo simulations to assess uncertainty and tail risk

Responsibilities

Discover and define new quantitative risk signals where no established metrics exist, using subject matter expertise to surface early, weak, or unconventional indicators
Translate complex trust and safety challenges into measurable signals that can be tracked and stress-tested over time
Develop upstream early-warning and signal frameworks that inform downstream detection and mitigation efforts
Analyze risk trends to assess the underlying drivers and causal factors behind those changes
Conduct data mining and statistical modeling to understand how risks originate, evolve, and propagate across systems
Design adversarial scenarios, and quantitative stress tests to assess exposure, coverage gaps, and vulnerabilities
Produce clear data-driven briefs to support risk prioritization, contingency planning, and strategic risk products across teams