Director, AI Alignment and Interpretability (Remote)

What you'll do

• Own the alignment and interpretability research agenda for security-domain AI systems.
• Develop methods to detect misuse signals in model internals by probing latent representations and analyzing activations.
• Create evaluation frameworks with behavioral constraints and training interventions to ensure models operate within intended bounds.
• Publish original research on interpretability for security-specialized models to advance the field.
• Recruit, develop, and retain a lean team of research scientists while remaining an active technical contributor.

What we're looking for

• MS or PhD in machine learning, computer science, or related field with research depth in interpretability.
• 8+ years of experience in ML research or engineering, including direct work on large language model alignment.
• Hands-on expertise with mechanistic interpretability methods applied to real models.
• Experience designing and running rigorous alignment evaluations for AI safety claims.
• Track record of leading and growing researchers while remaining an active technical contributor.
• Background in offensive security, vulnerability research, or adversarial ML to understand misuse potential.