💼 About This Role

You'll own OpenAI's holistic approach to identifying and forecasting frontier risks from AI systems. You'll ensure evaluation frameworks, safeguards, and taxonomies are robust and forward-looking. This role shapes the rationale for prioritizing catastrophic risk mitigation across technical, governance, and policy domains.

🎯 What You'll Do

• Develop threat models across misuse areas like bio, cyber, and attack planning.
• Model alignment risks such as loss of control and self-improvement.
• Forecast risks using technical foresight and adversarial simulation.
• Translate threat models into actionable mitigation designs with technical partners.

📋 Requirements

• Deep experience in threat modeling or adversarial thinking.
• Understanding of frontier AI risks and AI alignment literature.
• Knowledge of how AI evaluations connect to capability testing and safeguards.

✨ Nice to Have

• Experience in security, national security, or safety domains.
• Ability to communicate complex risks to non-technical audiences.
• Systems thinking with anticipation of second-order risks.

🎁 Benefits & Perks

• 💰 Competitive compensation with equity.
• 🏖️ Flexible PTO (not explicitly listed, but implied in tech culture).
• 🧠 Work on cutting-edge AI safety.