I am a PhD candidate at the University of Southern California's (USC) Information Sciences Institute (ISI), advised by Jonathan May. Previously, I was a programmer analyst at ISI and did my undergraduate studies at Hong Kong University of Science and Technology (HKUST) with a major in Computer Science, where I first worked on NLP with Pascale Fung. I've interned at Meta in 2021 and 2022 and at Amazon AGI in 2023 and 2024.

My research focuses on building human-centric AI systems that effectively collaborate with us by interactively understanding and fulfilling our intended goals, which are often communicated unclearly or without necessary context. I am interested in creating virtual environments with user models and task-specific user interfaces to simulate noisy human-AI interactions of complex real-world challenges that are impractical to study otherwise. With high-fidelity simulations, I explore generating high-quality synthetic data at scale and designing novel reward functions for reinforcement learning and rejection sampling to post-train models. A central theme of my work is human-AI context synchronization, expanding the scope of context shared by humans and AI about each other's knowledge, goals, and capabilities to balance AI utility with human agency and privacy.