Researcher, Safety Training
OpenAI · London, United Kingdom; San Francisco, United States · Full-time
Team: Research
Posted: 2023-05-25
Pay: USD 380000 – USD 500000 per year
The Safety Training Researcher develops and implements post-training methods to improve the safety, reliability, and alignment of AI models. This role involves conducting research on techniques like RLHF and DPO, building evaluation infrastructure, and collaborating cross-functionally to mitigate risks in production systems. Candidates should possess a PhD or equivalent experience in machine learning, with hands-on expertise in fine-tuning and post-training language or multimodal models.
Skills & tags: machine learning, rlhf, dpo, ppo, ai safety, model alignment, adversarial training, synthetic data, evaluation infrastructure, natural language processing
View and apply at the original posting
Postings belong to their employers; this page links to the source.