Yong Zheng-Xin announced on X that he has started full-time at OpenAI as a Member of Technical Staff, working on RSI preparedness and safety.
Yong studies how advanced AI systems can become misaligned during training or deployment, and how researchers can detect dangerous behavior before release. His recent work includes evaluating frontier-model risks, monitoring hidden reasoning, and studying why safety protections can fail when models operate across different languages.
The position follows an Astra Fellowship at OpenAI from January through June 2026 and the completion of his computer-science PhD at Brown University. His earlier multilingual-model work received best-paper awards at ACL 2024 and the NeurIPS 2023 Socially Responsible Language Modeling workshop.