Agent Post-Training Researcher

OpenAI; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

OpenAI is seeking a highly capable professional for the Agent Post-Training team to advance frontier agent capabilities. You will design experiments, improve training stacks, and build evals that reveal model failures, translating findings into product improvements.

The role spans research, engineering, data, and product work, with opportunities to own a direction, optimize pipelines, and drive robust deployment in real-world settings. Location: SF or London, hybrid option.

Qualifications

• Strong fundamentals in ML, software, systems, or statistics.
• Hands-on experience with LLMs, RL, RLHF/RLAIF, post-training, evals, graders, and synthetic data.
• Ability to move from vague problem to concrete experiments and production-ready results.

Responsibilities

• Design and run experiments to improve agentic model behavior.
• Own end-to-end improvements to post-training stack.
• Build evals and environments to expose model failures.
• Collaborate with product teams to translate signals into model improvements.

Skills

Machine learning

Software engineering

RLHF

Post-training

Eval design

Tools

Python

TensorFlow

PyTorch

Job description

OpenAI is seeking a highly capable professional for the Agent Post-Training team to advance frontier agent capabilities. You will design experiments, improve training stacks, and build evals that reveal model failures, translating findings into product improvements.

The role spans research, engineering, data, and product work, with opportunities to own a direction, optimize pipelines, and drive robust deployment in real-world settings. Location: SF or London, hybrid option.