Mentor, PRISM 2026

• Guide a team of 4 AI safety researchers from research conception to conference submission.
• Review peer-vetted work at weekly milestones, focusing on strategic direction and high-level feedback.
• Shape research direction, methodology, and experimental design in AI safety domains like interpretability and alignment.
• Attend 9 focused mentor sessions across 16 weeks (June-September 2026) totalling approximately 14 hours.

This role directly advances AI safety research by mentoring emerging researchers, helping produce rigorous, peer-vetted work that contributes to safer AI development.

• Expertise in AI safety research, particularly interpretability and alignment.
• Experience guiding research teams and providing strategic feedback.
• Ability to review and shape experimental design and methodology.
• Commitment to attend 9 mentor sessions over 16 weeks.

Compensation not specified. Role offers the opportunity to shape cutting-edge AI safety research and mentor the next generation of researchers.

Peer-Vetted Research Initiative for Safety Methodologies (PRISM) is an organization focused on advancing AI safety through rigorous, peer-reviewed research methodologies.