💼 About This Role

You'll post-train frontier models to autonomously run parts of the scientific discovery pipeline, including generating hypotheses and designing experiments. Your work will directly enable novel scientific discoveries. You'll collaborate with world-class experts in physical sciences.

🎯 What You'll Do

• Post-train frontier models for scientific discovery pipeline
• Create high-quality evaluation and training tasks
• Scale up RL environments for LLMs
• Design creative reward functions and run RL runs

📋 Requirements

• Experience with RL environments for LLMs
• Experience creating high-quality evals for frontier models
• Skill in training datasets and reward functions with LLMs
• Ability to collaborate with domain experts on evaluation criteria