Research Engineer, Multimodal Safety

Job Description

Job Title: Research Engineer, Multimodal Safety. Location: San Francisco, CA (Hybrid – 3 days in office)

Responsibilities:

• Build and maintain evaluation pipelines to assess AI risks, particularly for multimodal models.
• Implement risk mitigation techniques, including safety reward models and reinforcement learning.
• Develop and improve multimodal moderation models to prevent AI misuse;
• Collaborate with internal teams to update content policies and prevent harmful behavior;
• Participate in internal and external red-teaming efforts to evaluate model robustness.
• Write high-quality, efficient, and maintainable code for evaluation libraries.

Qualifications:

• Strong experience with multimodal research and post-training techniques.
• Proficiency in Python and model evaluation frameworks.
• Experience with large data pipelines and human data integration;
• Ability to communicate effectively with both technical and non-technical stakeholders.
• Previous experience working in complex technical environments is a plus;
• Passion for AI safety and innovation in the field of AI deployment.

Benefits & Instructions:

• Competitive salary range of $310K – $460K, plus equity options.
• Relocation assistance for new employees.
• OpenAI is committed to a diverse and inclusive workforce;
• Equal opportunity employer; we encourage applicants from all backgrounds to apply;
• Reasonable accommodations are available for applicants with disabilities.