Data Scientist, Cybersecurity

Overview

OpenAI is hiring a senior data scientist to help define and measure cybersecurity in the era of AI agents. This role sits within the Agentic Data Science team and partners closely with Security, Cyber Product, Engineering, and Research to build measurement frameworks, improve security controls, and turn ambiguous security problems into actionable decisions.

You will help evaluate how AI systems behave in security-sensitive settings, assess whether controls reduce risk without adding unnecessary friction, and build the analytical foundation for security and product strategy. The work spans internal security operations and customer-facing AI-powered cybersecurity products.

What You'll Do

• Define metrics and evaluation frameworks for AI-agent security, including control coverage, agent behavior, sensitive actions, access patterns, detection quality, and emerging risks.
• Measure the impact of security controls and quantify tradeoffs such as false positives, blocked actions, escalations, approval delays, and recovery paths.
• Improve instrumentation, connect fragmented telemetry, establish trusted datasets, and identify data-quality and coverage gaps.
• Analyze signals of anomalous behavior, risky access, sensitive-data exposure, and other security-relevant activity.
• Evaluate whether interventions improve detection quality, response times, and real-world security outcomes.
• Assess AI-powered cybersecurity products and how well they identify security issues, support workflows, and create customer value.
• Develop evaluation systems for security findings, including accuracy, severity, actionability, duplication, resolution, and downstream impact.
• Measure the full security workflow from discovery and investigation through validation, prioritization, and remediation.
• Design rigorous experimentation strategies using controlled experiments, staged rollouts, and observational analysis.
• Translate analysis into security and product strategy for technical partners and senior leadership.
• Help build a new security data science capability and operating rhythm across teams.

Requirements

• 5+ years of experience in data science, applied research, analytics, or a related quantitative field.
• Background in cybersecurity, trust and safety, fraud or abuse prevention, privacy, platform integrity, or another adversarial domain.
• Strong proficiency in SQL and Python.
• Experience working with complex datasets, incomplete instrumentation, and reproducible analytical workflows.
• Experience defining metrics and evaluation frameworks when ground truth is limited or outcomes are delayed.
• Strong judgment in experimentation, causal inference, and observational analysis.
• Ability to partner effectively across security, product, engineering, research, and leadership teams.
• Comfort operating independently and building structure in a domain without established standards.

Preferred Qualifications

• Experience with detection engineering, threat research, security operations, insider risk, identity and access management, or privacy-preserving security analytics.
• Familiarity with AI agents, large language models, model evaluations, automated code review, or AI-powered cybersecurity products.
• Experience evaluating security findings, vulnerability detection, remediation workflows, or developer-facing security tools.
• Experience balancing security effectiveness against user experience, including false positives and approval flows.
• Experience building automated monitoring, anomaly detection, production data assets, or systems that connect model outputs to real-world outcomes.
• Track record of building cross-functional measurement programs or new analytical capabilities from scratch.