Alexander Chien

Hi! I’m a 4th Computer Science undergraduate at UCLA. I work on multimodal and agentic AI, guided by a recurring question: what should a system do when it realizes it might be wrong?

I’ve done research in UCLA’s StarAI and NLP groups, where I was advised by Guy Van den Broeck, Ying-Nian Wu, and Kai-Wei Chang. My work has focused on structured reasoning and grounding, with an emphasis on understanding when language models drift or hallucinate, especially in settings where those failures carry real consequences. I want to build AI systems that do not just sound confident, but know when to slow down and revise their beliefs.

I’m also interested in creative and interactive uses of AI, including art, animation, and systems that respond to human intent, as another way of studying how models perceive, decide, and communicate. Teaching is part of that same thread. Through exploretech.la, I work with high school students encountering machine learning for the first time, and I have learned that ideas only stick when they survive real examples and honest questions.