Researcher, Misalignment Research

About the Role

The Safety Systems team is committed to building a culture of rigorous, impact-oriented safety work, and as a researcher, you will be expected to contribute to this effort. You will be working on complex problems that require creative and innovative solutions, and you will have the opportunity to collaborate with other researchers, engineers, and policymakers to drive adoption of safety evaluations across Openai.

What You Will Do

• Design and implement worst-case demonstrations to reveal how AI systems can go wrong
• Develop adversarial and system-level evaluations to measure dangerous capabilities and residual risks
• Create automated tools and infrastructure to scale automated red-teaming and stress testing
• Conduct research on failure modes of alignment techniques and propose improvements
• Publish influential internal or external papers that shift safety strategy or industry practice
• Partner with engineering, research, policy, and legal teams to integrate findings into product safeguards and governance processes
• Mentor engineers and researchers to foster a culture of rigorous, impact-oriented safety work
• Collaborate with other labs and researchers to advance the field of AI safety
• Develop and maintain expertise in AI red-teaming, security research, and adversarial ML
• Stay up-to-date with the latest developments in AI safety and misalignment research

What We Are Looking For

• 4+ years of experience in AI red-teaming, security research, adversarial ML, or related safety fields
• Strong research track record, including publications, open-source projects, or high-impact internal work
• Fluency in modern ML / AI techniques and comfort with hacking on large-scale codebases and evaluation infrastructure
• Ability to communicate clearly with both technical and non-technical audiences
• Experience with collaborating on cross-functional projects that span research, engineering, and policy
• Ph.D., master's degree, or equivalent experience in computer science, machine learning, security, or a related discipline
• Strong understanding of AI safety and misalignment risks
• Experience with designing and executing cutting-edge attacks and evaluations

Nice to Have

• Experience with automated red-teaming and stress testing
• Knowledge of AI safety frameworks and methodologies
• Familiarity with Openai's products and technologies
• Experience with collaborating with external researchers and labs
• Strong programming skills in languages such as Python, C++, or Java

Benefits and Perks

• Competitive salary and benefits package
• Opportunity to work on cutting-edge AI safety research
• Collaborative and dynamic work environment
• Flexible work arrangements, including remote work options
• Professional development opportunities, including conferences and training
• Access to cutting-edge technologies and tools
• Chance to contribute to the development of safe and responsible AI

• To stand out as a candidate, be prepared to demonstrate your skills in designing and executing cutting-edge attacks and evaluations.
• Make sure to highlight your experience with AI red-teaming, security research, and adversarial ML in your application.
• Show a strong understanding of AI safety and misalignment risks, and be prepared to discuss your ideas for mitigating these risks.
• Emphasize your ability to collaborate with cross-functional teams, including research, engineering, and policy.
• Be prepared to discuss your research experience and publications, and how they relate to the role.
• Don't be afraid to ask questions about the role and the company during the interview process, it shows your interest in the position.