Scientific AI Researcher

Job Summary

Building and post-training agents for scientific research, the full-time Scientific AI Researcher will focus on creating environments for literature search, code execution, and experimental design while collaborating with research partners in a remote setting.

Key responsibilities

• Build long-horizon environments for agents to search literature, query databases, and run experiments within a defined budget
• Transform research partners' workflows and datasets into trainable tasks and evaluations
• Design verification methods for correctness and reproducibility, ensuring agents utilize evidence appropriately

Required qualifications

• Depth in reinforcement learning (RL), agents, or post-training systems with the ability to acquire scientific context
• Hands-on experience in building and debugging model-training or evaluation loops
• Strong Python programming skills to create artifacts for research and engineering use
• Demonstrated experimental judgment with experience in baseline evaluation and failure attribution
• Comfort with ambiguous problems and ability to keep partners informed during resolution