Researcher, Alignment Science

📋 Description

• Design and run alignment experiments focused on intent following, honesty, calibration, and robustness.
• Train and evaluate models using reinforcement learning, and other empirical ML methods.
• Develop evaluations for failure modes such as hallucination, instruction-following failures, reward hacking, covert actions, and scheming.
• Study methods that encourage models to verify their behavior and report shortcomings honestly, including confession-style training objectives.
• Build monitoring and inference-time interventions that ensure compliant behavior or surface model issues to users or downstream systems.
• Investigate how alignment methods scale with model capability, compute, data, context length, action length, and adversarial pressure.
• Integrate successful techniques into model training and deployment workflows.
• Produce externally publishable research when results advance the broader science of alignment.
• Collaborate with researchers and engineers across post-training, RL, evaluations, safety, and product-facing teams.

🎯 Requirements

• Hands-on experience training, evaluating, or debugging large ML models (LLMs).
• Strong Python and PyTorch engineering skills.
• Mathematical rigor and quantitative thinking.
• Experience with RL, post-training, or scalable ML research.
• Ability to work independently with minimal day-to-day guidance.
• Thrives in fast-paced, collaborative research environments.
• Strong track record in rigorous problem solving (competition, systems, etc.).
• Commitment to trustworthy, honest AI in high-stakes settings.
• Motivated to advance alignment methods that can be tested and deployed.

🎁 Benefits

• Hybrid work model: 3 days in office per week.
• Relocation assistance for new employees.
• Exceptional remote candidates considered.