Researcher, Agentic Post-Training

📋 Description

• Own end-to-end research and engineering projects for post-training OpenAI's agentic models.
• Decide with partner teams which integrations are ready for major model runs.
• Develop improvements across factuality, following, tool calling, and multi-agent behavior.
• Build and improve training, evaluation, grading, and data infra for large-scale RL runs.
• Create evals and diagnostics to assess model readiness.
• Improve feedback from product usage into post-training via implicit user signals.

🎯 Requirements

• Have strong ML fundamentals and hands-on experience with LLMs, RL, RLHF, post-training, or evals.
• Are a strong engineer who can move quickly in complex systems and make pragmatic decisions.
• Can own ambiguous problems end-to-end without needing a tightly specified roadmap.
• Care about impact over method and can do unglamorous, high-impact work.
• Have excellent taste in model behavior and reason about what good looks like.
• Are comfortable working across research, infra, data, evals, and product.