Researcher, Agent Safety, Training and Evaluations
About this role
The Agent Safety team at OpenAI builds systems and training methods to ensure that capable AI agents act safely, exercise sound judgment, and remain aligned with user intent. Professionals in this group focus on reducing severe unintended outcomes while maintaining useful agent autonomy through targeted training, rigorous measurements, and reliable oversight mechanisms.
Team members design training environments and data pipelines, build scalable evaluation metrics, and develop mitigation strategies for emerging risks. Candidates who possess strong research engineering or quantitative backgrounds and a pragmatic drive toward practical safety problems will find themselves well-suited for the position, even without prior alignment experience.
What you will do
- Train and evaluate frontier models to minimize harmful or misaligned agent actions by establishing clear hypotheses and executing them independently.
- Analyze real-world incidents to construct scalable measurement, data-processing, and evaluation systems that generate repeatable safety signals.
- Partner closely with post-training, capabilities, oversight, and pre-training teams to integrate research-backed mitigations into large-scale training and agent systems.
What you need
- Demonstrated strength in research engineering, machine learning engineering, quantitative research, or applied model research with the ability to own ambiguous projects end to end.
- Strong technical execution spanning experimentation, data handling, evaluation, and infrastructure alongside robust intuition for modern frontier-model research.
- High motivation for agent safety and readiness to tackle urgent, practical problems regardless of prior work history outside the safety or alignment domain.
Benefits
- Offers Equity
- Relocation assistance to new employees
Questions candidates ask
Where is this role located and what is the work model?
This role is based in San Francisco, CA. We use a hybrid work model of 3 days in the office per week and offer relocation assistance to new employees.
Do I need prior safety or alignment experience?
You don't need prior safety or alignment experience, as we also welcome people that recently realized that alignment and safety is a critical area to contribute to.
What does the compensation package include?
The salary ranges from $380K to $500K and includes offered equity.