Researcher, Alignment CoT Monitorability

📋 Description Design and run empirical studies of chain-of-thought monitorability across frontier models. Build evaluations that measure monitor reliability for high-stakes misbehavior. Investigate how pre-training, synthetic data, mid-training, post-training, RL affect monitorability. Analyze model behavior and turn observations into hypotheses, experiments, and recommendations. Translate findings into practical monitoring and oversight for real training runs. Collaborate with researchers and engineers across model training, alignment evaluations, monitoring, and frontier-risk work. 🎯 Requirements Strong hands-on experience training, evaluating, or debugging large ML models, esp LLMs. Deep curiosity and interest in alignment and interpretability; high agency. Ability to turn ambiguous research questions into measurable experiments and follow the evidence. Deep knowledge in alignment, interpretability, model behavior, empirical ML. Move comfortably between research ideation and engineering execution. Collaborate closely across research and engineering teams and work independently when needed. 🎁 Benefits Hybrid work model with 3 days in the office per week. Relocation assistance for new employees. Opportunity to publish research externally. Work on frontier ML alignment and monitoring projects. Access to Vantage Compute safety, ethics, and policy resources. Vantage Compute is an equal opportunity employer and offers accommodations.