I am a third-year Ph.D. candidate at KAIST advised by Professor Joon Son Chung, and I earned my M.S. from KAIST. Since June 2026, I have also been a Ph.D. Research Intern with the Qualcomm Speech Team in South Korea. My research focuses on personalized voice agents that understand the person they are speaking with and tailor their interactions to that person.
I envision a future in which voice-first AI agents and physical AI become part of everyday life through robots, smart glasses, and other platforms that enable interaction without screens or manual controls. In these screenless settings, voice can be the most natural, and sometimes the only, way to communicate. As people entrust agents with increasingly personal aspects of their lives and work, I believe that understanding who is speaking is as important as understanding what is said. The same words may call for different responses depending on the speaker, their relationship to the agent, and the situation. I aim to build agents that interpret each participant's speech appropriately while keeping track of the requests and conversational context of the user they are assisting.
My research at KAIST has evolved from robust speech and speaker representation learning to diffusion-based modeling and audio-text alignment. Building on this foundation, I now explore speaker-aware Audio-LLMs that connect speaker understanding with language-based reasoning.
At Qualcomm, I am researching streaming speaker-aware Speech LLMs. I also work on full-duplex Speech LLMs that understand and adapt to their conversation partners. I am particularly interested in agents that listen while speaking and respond to incoming speech according to who is speaking and the conversational context. My goal is for speaker understanding to shape both response content and decisions about when to listen, speak, and respond.
I welcome conversations and collaborations on speech and speaker understanding, personalized voice agents, and voice interaction with robots and wearables. Please feel free to get in touch.
Advisor: Joon Son Chung (Multimodal AI Lab)
Advisor: Joon Son Chung (Multimodal AI Lab)
