Research Engineer, Human Data

The Human Data team turns human feedback into reliable signals for training and evaluation. We design and run end-to-end programs that capture the depth of human intent behind everyday and high-stakes uses of our models. Our remit spans bespoke data campaigns, scalable synthetic data generation, and product-embedded signals. We partner closely across all research teams to translate these signals into training datasets, novel evaluations, and feedback loops that push the frontier of our models and advance their applications.

About the Role

As a Research Engineer on Human Data, your mission will be to invent and ship novel systems that maximize the utility of human feedback at the capability and deployment frontiers. You’ll create new ways to capture and weight explicit feedback, decode implicit signals from our products, and extract deeper insights from every bit of supervision. We’re looking for people who pair first-principles curiosity with strong engineering ability and bias to impact.

This role is based in San Francisco, CA. We use a hybrid work model of at least 3 days in the office per week and offer relocation assistance to new employees.

In this role, you will:

• Partner with researchers and the Human Data product manager to design, derisk, and implement 0-to-1 innovations in all aspects of human feedback.

• Uncover new insights on our AI trainers and users alike, and apply these findings to OpenAI’s research goals.

• Measure the efficacy of our existing feedback methods, and experiment to increase that efficacy.

• Design methods that precisely capture human feedback across subjective, context-dependent goals and evolve with how people actually use our models.

• Ensure our evaluations accurately track real-world utility and evolve with how people use our models and applications.

• Run online and offline experiments that quantify your downstream causal impact on model performance.