Career Profile
I’m a Ph.D. student in the Music and Audio Computing Lab at KAIST, advised by Prof. Juhan Nam. My research focuses on representation learning and generative modeling for speech and music, with an emphasis on audio tokenization, disentangled representations, and controllable synthesis. My work spans neural speech codecs, music generation and transcription, and the evaluation of audio-language models.
Outside research, I enjoy playing acoustic guitar and writing songs and lyrics. Some of my favorite artists are Radiohead, Tatsuro Yamashita, the Velvet Underground, and Byeong-woo Lee.
Education
Music and Audio Computing Lab (Advisor - Juhan Nam)
Music and Audio Computing Lab (Advisor - Juhan Nam)
Experiences
Publications
The Fortieth Annual Conference on Neural Information Processing Systems (NeurIPS 2026), Evaluations & Datasets Track
A benchmark for evaluating audio-language models on expert vocal coaching feedback for singing.
SDP-Codec: A Speaker-Decoupled Speech Codec with Pitch Injection for Low-Bitrate Coding and Zero-Shot Voice Conversion
[code]
[demo webpage]
[demo video]
The 27th Annual Conference of the International Speech Communication Association (Interspeech), 2026
Neural speech codec for disentangled speech representation learning, low-bitrate coding, and zero-shot voice conversion.
D3PIA: A Discrete Denoising Diffusion Model for Piano Accompaniment Generation From Lead Sheet
[code]
[demo]
Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2026
A lightweight discrete diffusion model for lead-sheet-conditioned symbolic music generation.
Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2025
Discrete diffusion-based refinement model for structured piano transcription.
Proceedings of the 27th International Conference on Digital Audio Effects (DAFx24), 2024
Best Student Paper Award
Controllable neural audio effect modeling.
Expressive Acoustic Guitar Sound Synthesis with an Instrument-Specific Input Representation and Diffusion Outpainting
[demo]
Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2024
Controllable acoustic instrument synthesis using instrument-specific representations and diffusion outpainting.
J. Kor. Powd. Met. Inst., 20.5 (2018) 376-381
Projects
ToplAIner - Acoustic Guitar Conditioned Singing Voice Generation [Poster]
- 2023 Fall KAIST GCT731 Topics in Music Technology Final Project
Academic Advising
Teaching Activities
Skills
Python & PyTorch·
Deep Learning & Generative Models·
Audio, Speech & Music Signal Processing·
Representation Learning & Audio Tokenization·
Diffusion Models, Flow-based Models & Neural Codecs·
Experiment Management, Model Training & Evaluation·
Linux Server Administration, Docker, SLURM & Git