Researcher, Alignment CoT Monitorability (AI)

TL;DR

Researcher, Alignment CoT Monitorability (AI Alignment/ML): Studying whether the chain-of-thought of frontier reasoning models is monitorable to support scalable oversight with an accent on measuring monitorability and investigating training mechanisms. Focus on designing empirical studies of model behavior, building evaluations for high-stakes misbehavior, and translating research into practical training recommendations.

Location: Based in San Francisco, CA (Hybrid: 3 days in office). Relocation assistance provided.

Salary: $250K – $445K + Equity

Company

OpenAI is an AI research and deployment company dedicated to ensuring that general-purpose artificial intelligence benefits all of humanity.

What you will do

• Design and run empirical studies of chain-of-thought monitorability across frontier reasoning models and training settings.
• Build evaluations that measure whether monitors can reliably predict properties of interest, including high-stakes misbehavior.
• Investigate how interventions such as pre-training, synthetic data, RL, and post-training influence reasoning legibility.
• Analyze model behavior and translate observations into hypotheses, experiments, and practical oversight recommendations.
• Collaborate with researchers and engineers across model training, alignment evaluations, and frontier-risk research.
• Produce externally publishable research to advance the broader science of alignment.

Requirements

• Strong hands-on experience training, evaluating, or debugging large ML models, especially LLMs.
• Depth in alignment, interpretability, model behavior, empirical ML, or adjacent research.
• Ability to transform ambiguous research questions into concrete, measurable experimental setups.
• Capacity to move comfortably between research ideation and engineering execution.
• Must be based in or be able to relocate to San Francisco, CA.

Culture & Benefits

• Hybrid work model requiring 3 days of office attendance per week.
• Relocation assistance provided for new employees.
• Opportunity to apply monitoring methods to OpenAI's largest RL training runs.
• Work in a high-agency environment dedicated to making AI systems more monitorable, trustworthy, and safe.