🛠 Tech Stack

💼 About This Role

You'll own measurement and quantitative analysis for severe usage harm verticals at OpenAI, developing AI-first methods for prevalence estimation. You'll build metrics, dashboards, and insights that directly improve safety systems and influence roadmaps. This role is based in London and involves sensitive content, requiring deep trust and safety expertise.

🎯 What You'll Do

• Own measurement and metrics for multiple harm verticals
• Develop AI-first prevalence measurement methods
• Build dashboards and metrics for goal-setting and A/B tests
• Conduct analyses to inform detection and enforcement improvements

📋 Requirements

• Senior-level data science experience with trust and safety
• Deep statistics skills in sampling and prevalence estimation
• Experience with sensitive harm areas like child safety or violence
• Proficiency in Python/R and SQL

✨ Nice to Have

• Experience with AI harms or leveraging AI for measurement
• Experience with off-platform or non-standard datasets

🎁 Benefits & Perks

• 🏖️ Flexible PTO
• 🏥 Comprehensive health insurance
• 📈 Equity grants
• 🧑‍💻 Home office budget