Job description
OpenAI • San Francisco Bay Area
San Francisco Bay Area
$250,000 - $445,000
- In this role, you'll design alignment experiments ensuring AI models follow user intent, remain honest, and transparently report limitations.
- Train and evaluate large language models using reinforcement learning and empirical methods.
- Develop evaluations for model failure modes including hallucination, instruction-following failures, and scheming.
- Build monitoring and inference-time interventions to ensure model compliance and surface issues.
- Integrate successful alignment techniques into model training and deployment workflows, producing publishable research.
OpenAI is a frontier AI research and product company, with teams working on alignment, policy, and security. You can read concerns about doing harm by working at a frontier AI company in our career review on the topic, including concerns about OpenAI in particular.
Our Take On This Role: We have concerns about OpenAI's track record on safety and responsible development and do not recommend almost any roles at OpenAI. Nonetheless, it is possible that OpenAI will create AGI in the next decade, in which case safety and security work at the company could be extremely important. If you receive a job offer from OpenAI, consider contacting us for career advice.