New opportunity

Agent Post-Training, Personality

OpenAI · San Francisco, United States

About this role

Join the Agent Post-Training Personality team to study and shape agent behavior, translating insights into evals, training data, reward signals, and model improvements; collaborate with product teams, human experts, and researchers to carry improvements through training and into deployed agents.

Skills for this role

LLMsPost-trainingRLHFReward ModelingReinforcement LearningEvalsSynthetic DataPretraining DataTraining DataProduction ML SystemsBehavioral ResearchProduct ThinkingHCIStatisticsSoftware EngineeringGrader DevelopmentMulti-agent CoordinationLong-horizon ExecutionInstruction FollowingCalibrated ReasoningFactualityPreference DataHuman-Data CollaborationTraining Pipelines