Curriculum Vitae
I’m a researcher in voice and audio AI and co-founder and technical lead at Studio Cucurbits. I work across machine learning, signal processing and audio software, turning research into tools for speech, music and artistic production.
Research Interests
AI for voice, music, and sound:
- Controllable and interpretable generative models
- Self-supervised representation learning
- Neural audio/speech codecs
- Music applications: MIR, singing voice synthesis, music generation models
- Speech applications: TTS, VC, ASR
Open to new challenges related to artistic media.
Experience
Studio Cucurbits. • Korea
- Software engineering and QA for an AI-based pipeline that develops a suite of VST plugins
- Release pipeline for commercial audio effects (GitHub CI, signed builds); first product on sale since 2026
- AWS/Cloudflare infrastructure: CloudFront site, Lambda licensing, DNS, R2 delivery; Paddle shop
- Four-language phonetic relay prototype combining speech recognition, language-model interpretation, speech synthesis and OSC control for an interactive installation
Supertone Inc. • Korea
- Developed, trained, and packaged the voice backbone NANSY (~400M parameters; ~1B pipeline)
- Self-supervised pitch with subharmonics and jitter/shimmer; hybrid neural/Viterbi decoding
- Reduced the model from ~400M to ~150M parameters; distilled real-time variants
- LM-based TTS: variable-frame-rate tokenizer; Qwen3-1.5B fine-tuning; training on up to 16 H100s
- Voice-editing tool for B2B clients (pitch, subharmonics, noise); ISMIR 2025 late-breaking demo
IRCAM • France
- Deep learning for voice: neural vocoding, representation learning, voice attribute transformation
- Bespoke voice transformations for film, music, and art
- Turned research into voice-transformation software and a VST plugin (CIRCE)
- Postdoctoral research: September 2023 - March 2024; doctoral research: May 2019 - September 2023; internship: July 2018 - January 2019
Siemens AG • Germany
TU Darmstadt, Signal Processing Group • Germany
Education
Sorbonne Université • France
TU Darmstadt • Germany
TU Darmstadt • Germany
Additional Education
TU Darmstadt, Akademie für Tonkunst • Germany
Tongji University • Shanghai, China
AC Davis High School • Yakima, WA, USA
Skills
Research: Experiment design and evaluation · scientific writing · publications at ICASSP, Interspeech, ISMIR
Machine learning: Generative modeling · self-supervised learning · neural codecs and vocoders · distillation
Domain knowledge: Signal processing · acoustics · voice science · music theory
Engineering: Python (PyTorch, NumPy) · AI-assisted engineering and QA · CI/CD · AWS · model deployment
Languages: German (native) · English (C2) · French (B2) · Polish (B1) · Chinese (A2)
Music: Classically trained singer · composition · live coding (Tidal Cycles) · piano
Scholarships
Activities
Talks
Code
Checkout my github page for code examples, though, unfortunately, most of my actual code cannot be made available.