Page Archive

PhD student at USC building LLMs and LLM agents that improve themselves, that humans can trust, and that are efficient to train and run.