Agent Post-Training, Personality

About the Position

Agent Post-Training, Personality

San Francisco

About the Team

The Agent Post-Training team at OpenAI is at the forefront of creating frontier agents that power our products like Codex, ChatGPT, and the API. We focus on developing persistent, proactive intelligence that can operate computers, collaborate with people and other agents, and expand the possibilities of what people and organizations can achieve. Our work encompasses coding, tool use, computer use, multi-agent coordination, long-horizon execution, factuality, instruction following, calibrated reasoning, and taste.

We build the data, environments, graders, training methods, and feedback loops that shape the capabilities and characteristics of our agents. These improvements are then integrated into major training runs and ultimately into the products used by people every day.

About the Role

As a member of the Agent Post-Training Personality team, you will play a crucial role in making OpenAI’s agents exceptional collaborators. You will study and define what makes an agent thoughtful, clear, perceptive, appropriately proactive, and genuinely easy to work with. This involves translating those insights into evaluations, training data, reward signals, and model improvements.

Our concept of "personality" goes beyond writing style or general likability. It includes understanding user goals, communicating with good judgment, adapting to context, asking useful questions, handling disagreement honestly, and taking initiative at the right moments. The goal is to create a strong, tasteful default that can adapt to different people and situations.

This role combines behavioral research, product thinking, and research with a focus on communication and taste. You will collaborate with product teams, human experts, and researchers across post-training and pretraining to ensure that improvements are integrated into the full training stack and reach the models people use every day.

Responsibilities

• Develop a rigorous understanding of what makes an agent a great collaborator across various domains.
• Translate qualitative judgments about model behavior into concrete hypotheses, evaluations, graders, and training interventions.
• Study explicit and implicit user signals to understand which behaviors create trust, satisfaction, continued use, and successful outcomes.
• Work with human experts and trainers to produce high-quality, tasteful rollouts and preference data.
• Improve reward models and reinforcement learning objectives for model behaviors.
• Collaborate with pretraining and early-training teams on data mixtures, objectives, synthetic data, and other upstream choices.
• Build sustainable pipelines for updating older training data as our understanding of excellent model behavior evolves.
• Partner closely with product teams to turn consumer insights into model improvements and validate them in real workflows.
• Own projects end to end, from identifying behavioral failures through experimentation, training, evaluation, and launch.

Requirements

• Think instinctively from the user’s perspective and care deeply about how models feel to work with.
• Translate subjective product questions into falsifiable hypotheses and rigorous evaluations.
• Care about preserving individuality, adaptability, and behavioral diversity.
• Shape how frontier agents communicate, collaborate, and build trust with millions of people.
• Have strong technical foundations in machine learning, software engineering, statistics, behavioral science, HCI, or related fields.
• Have a strong taste for model behavior and can explain why certain responses feel thoughtful, natural, and useful.
• Have experience with LLMs, post-training, RL/RLHF, reward modeling, evaluations, synthetic data, pretraining data, or production ML systems.
• Are excited by ambiguous capability problems and enjoy working in environments where the solution may involve data, training, evaluations, product changes, or all of the above.
• Can work effectively with researchers, engineers, product teams, designers, domain experts, human-data teams, and safety boundaries.
• Like building load-bearing systems and processes when needed.

Nice to Have

• Research experience
• Strong communication skills
• Collaboration skills
• Project management experience
• Systems building experience

We Offer

• Opportunity to shape the future of AI agents
• Collaborative and innovative work environment
• Access to cutting-edge technology and resources

About the Company

OpenAI is an AI research and deployment company dedicated to ensuring that general-purpose artificial intelligence benefits all of humanity. We push the boundaries of AI capabilities and seek to safely deploy them through our products. We value diverse perspectives and experiences and are committed to providing reasonable accommodations to applicants with disabilities.

Join us in shaping the future of technology.