Researcher, Cyber Safety Training
OpenAI · San Francisco
Frontier labonsite$380k–$500k
At a glance
Highlights
- Work directly on large training runs shaping OpenAI's frontier models.
- Teach models to aid defenders while refusing harmful security assistance.
- Design realistic cyber safety evaluations to improve post-training methods.
What they offer
- Equity
Ideal candidate
Hands-on ML practitioner with practical security judgment who enjoys ambiguous cases and wants to bring cybersecurity expertise into frontier model training.
Requirements
- Trained or fine-tuned models and used experiment results to change next training run
- Practical security judgment from industry work, research, security teams, or CTF competitions
- Build and debug training experiment code from data prep to behavior validation
- Work through ambiguous cases with domain experts to define model responses
Lab profile
About OpenAI
Builds safe frontier AI models and products like ChatGPT for broad benefit.
What they build
- ChatGPT
- OpenAI API
- Frontier AI models
Known for
- Frontier-scale distributed ML training
- Massive GPU fleet and data center operations
- Exabyte-scale data lakes and streaming
- AI alignment and safety research
Key facts
- Operates one of the world's largest GPU fleets
Posted · Last seen