OpenAI
Agent Post-Training Research
$380K–$500K
San Francisco · Full-time · Posted Jun 25, 2026
About the job
Improve the capabilities, reliability, and product fit of agentic models through research, post-training, evaluation, engineering, and product integration.
What you'll do
- Design and run experiments to improve agent behavior and own improvements across the post-training stack.
- Build evaluations and environments that reveal model failures, and turn findings into training data, fixes, or research directions.
- Partner across product, research, and infrastructure teams to integrate, measure, and launch model improvements.
What they ask for
- Strong technical fundamentals in machine learning, software engineering, systems, statistics, or a related field.
- Hands-on experience with one or more of LLMs, RL, RLHF/RLAIF, post-training, evals, graders, synthetic data, model training, coding agents, tool-using agents, or production ML systems.
Pay
$380K – $500K (base salary)
Tools
LLMs, RL, RLHF/RLAIF
AI focus
Post-training and alignment research for agentic models, including training, evaluations, tool and computer use, multi-agent coordination, and shipping improvements into products.
Model-written summary · the original posting is authoritative