Agent Post-Training Research Engineer

OpenAI; California (MO); Hybrid; USD 190,000 - 230,000; Full time

Job summary

OpenAI is seeking a member of the Agent Post-Training team to improve the capabilities, reliability, and product fit of agentic models.

You might own a research direction, build infrastructure for large training runs, create evals, or drive a capability from idea through experimentation, integration, and launch.

This is a high‑agency role that may be based in San Francisco or London, with opportunities to ship improvements into OpenAI products used by developers and everyday users.

Qualifications

• Strong technical fundamentals in machine learning, software engineering, systems, or statistics.
• Hands-on experience with LLMs, RL, RLHF/RLAIF, post-training, evals, graders, synthetic data, model training, coding agents, tool-using agents, or production ML systems.
• Excited by open‑ended problems where research taste + engineering execution are needed.
• Care about product impact and model behavior, not just benchmark movement.

Responsibilities

• Design and run experiments that improve agentic model behavior across coding, tool use, function calling, computer use, and multi‑agent collaboration.
• Own end-to-end improvements to the post‑training stack, including RL, data pipelines, graders, reward signals, evals, diagnostics, and model‑behavior analysis.
• Build evals and environments that expose the next set of model failures and turn those into training data or product fixes.
• Partner with Codex, API/platform, and ChatGPT product teams to translate product signal into model improvements.
• Work on early‑training and alignment interventions, including data mixtures, objectives, synthetic data, and eval loops.

Skills

Machine learning fundamentals

Software engineering

Systems

Statistics

Tools

RLHF/RLAIF

Post-training

Evals

Graders

Synthetic data

Job description

OpenAI is seeking a member of the Agent Post-Training team to improve the capabilities, reliability, and product fit of agentic models.

You might own a research direction, build infrastructure for large training runs, create evals, or drive a capability from idea through experimentation, integration, and launch.

This is a high‑agency role that may be based in San Francisco or London, with opportunities to ship improvements into OpenAI products used by developers and everyday users.