Open to opportunities

쿠룰루룰 쿠

@0006124

Message

I build Korean and English TTS, speech, and NLP systems with machine learning.

South Korea

Message

I'm building TTS systems at Hana Institute of Technology, following machine learning research engineering work at MediaZen across Korean and English speech products. I've developed real-time Korean and English TTS, multi-speaker TTS, voice cloning, and fast multi-channelled TTS.

My work also spans Korean NLP, ASR, speech separation, and data systems, including customized Korean ALBERT models, Hyundai Motor Company NLU fine-tuning, TorchScript deployment for embedded systems, and a YouTube dictation data-collection platform. I hold a PhD in Phonetics from Korea University and work with TensorFlow, PyTorch, Python, Bash, Kaldi, MySQL, and Tornado.

Experience

Work history, roles, and key accomplishments

Current

머신러닝 엔지니어 (TTS)

Current

Hana Institute of Technology (하나금융융합기술원)

Oct 2024 - Present (1 year 11 months)

Worked as 머신러닝 엔지니어 (TTS) at Hana Institute of Technology (하나금융융합기술원).

ai팀 팀장

아이스크림아트

Nov 2023 - Oct 2024 (11 months)

Worked as ai팀 팀장 at 아이스크림아트.

Machine Learning Research Engineer

MediaZen-미디어젠

Jun 2019 - Oct 2023 (4 years 4 months)

On Natural Language Processing
- Customized Korean albert-base model and pre-trained with low resources (distributed learning & accumulated gradients)
- Fine-tuned albert base for Hyundai Motor Company NLU task (intent classification & slot tagging)
- Worked as a Hackathon (by National Information Society Agency, NIA (한국정보화진흥원)) mentor for emotion detection in dialogues

On Text to Speech System
-

Machine Learning Research Engineer

MediaZen-미디어젠

Aug 2017 - May 2019 (1 year 9 months)

On Natural Language Processing (NLP)
- Developed written-to-spoken language conversion system for Korean language

On Text to Speech System (TTS)
- Developed real-time Korean TTS
- Developed real-time English TTS
- Developed voice cloning system using neural network

On Automatic Speech Recognition (ASR)
- Supported developing Kaldi based ASR system optimal for specific domains
- Developed web cra

Education

Degrees, certifications, and relevant coursework

Coursera

Foundations of Project Management

Issued Jan 2023

Coursera

Object-Oriented Data Structures in C++

Issued Mar 2023

Korea University

Doctor of Philosophy - PhD, Phonetics

2020 - 2022

Korea University

Master's degree, Phonetics

2017 - 2020

Lectured in Hankuk University of Foreign Studies (한국외대) on Kaldi for basic ASR to programming beginners (2018.1)

On M/L
- Studied Text to Speech (TTS) System using M/L
- Studied Voice Cloning using M/L
- Studied Speaker Verification system using M/L

On Programming
- Studied BASH scripting
- Studied Python
- Studied tensorflow
- Studied pytorch
- Studied HTML/CSS/javascript and tornado python web