Julia Chae

I am a EECS PhD student at MIT CSAIL advised by Sara Beery. My PhD has been generously supported by the NSERC-PGS-D and MIT Andrew and Erna Viterbi Graduate Fellowship.

My research focuses on bridging general-purpose models with expert-level intelligence. I am interested in how evolving domain knowledge can be integrated into large AI models — through curated synthetic data, targeted self-supervision, and expert feedback — while preserving their adaptability and generalization. More broadly, I aim to scientifically understand how to overcome real-world data gaps, guide scalable and robust specialization, and design models that, like humans, refine their expertise in alignment with the structure of complex real-world environments.

I’m particularly interested in applications to ecology and biodiversity monitoring - previously, my work has focused on robotics applications.

Prior to my PhD, I received by BASc at the University of Toronto in Engineering Science, majoring in Robotics with a minor in Machine Intelligence. At UofT, I worked with Prof. Sanja Fidler (affiliations: UofT Machine Learning, Vector Institute, NVIDIA) on self-supervised representation learning, and with Prof. Florian Shkurti (affiliations: Vector Institute, UofT Robotics Institute), Prof. Steven Waslander (affiliations: UTIAS, U of T Robotics Institute) on robot learning and perception. Please see my CV below for more details.

If you would like to chat, feel free to reach out to me via one of the contact information below.