Ryuichi Yamamoto
Senior Research Scientist at LY Corporation
Nagoya, Japan
Summary
🤩
Rockstar🎓
Top SchoolRyuichi Yamamoto is a senior research scientist and technical lead for text-to-speech teams with 13 years of experience building production-ready speech synthesis and voice conversion systems. He combines deep ML expertise—authoring and refining PyTorch implementations of DeepVoice3 and WaveNet vocoders—with practical backend and DSP contributions to major open-source projects like librosa and ESPnet. His work spans both model architecture and systems engineering: improving ISTFT reconstruction, implementing efficient non-autoregressive waveform generators, and hardening cross-platform build systems. Ryuichi has led TTS efforts at LINE and LY Corporation and has published/demoed fast parallel TTS research accepted to Interspeech, showing a strong track record of turning research into deployable demos. Based in Nagoya, he brings a rare blend of academic rigor (PhD-level study) and hands-on engineering, frequently contributing build fixes and platform integrations that quietly improve developer experience across projects. Colleagues rely on him for robust, production-minded speech systems that bridge cutting-edge models and real-world constraints.
12 years of coding experience
10 years of employment as a software developer
Master of Engineering (M.Eng.), Computer Science, Master of Engineering (M.Eng.), Computer Science at Nagoya Institute of TechnologyDoctor of Philosophy - PhD, Informatics, Drop out in Oct. 2025, Doctor of Philosophy - PhD, Informatics, Drop out in Oct. 2025 at Nagoya University
Japanese, Chinese, Korean
Stackoverflow
21reputation
1kreached
1answer
0questions
Github Skills (41)
speech-to-text10
pytorch10
opencv10
scipy10
convolutional-neural-networks10
c-language10
speech-processing10
python10
signal-processing10
testing10
machine-learning10
text-to-speech10
speech-synthesis10
360-video10
audio-analysis10
Programming languages (10)
Github contributions (5)
r9y9/wavenet_vocoder
Dec 2017 - Nov 2020
WaveNet vocoder
Role in this project:
ML Engineer
ML EngineerContributions:1 release, 261 commits, 25 PRs in 2 years 10 months
Contributions summary:Ryuichi's commits primarily focus on implementing and refining a WaveNet vocoder for speech synthesis. They started with initialization and setup, including integrating deepvoice3_pytorch. The user then developed the core WaveNet architecture, including convolutional layers, gated linear units (GLU), and support for local and global conditioning. Further commits indicate the user implemented and tested incremental decoding techniques, along with modifications for a better softmax implementation, and added support for mixture of logistic distributions.
wavenet-vocoderwavenetspeech-enhancementvocoderspeech-processing
r9y9/deepvoice3_pytorch
Oct 2017 - Nov 2020
PyTorch implementation of convolutional neural networks-based text-to-speech synthesis models
Role in this project:ML Engineer
ML EngineerContributions:1 release, 2 reviews, 218 commits in 3 years
Contributions summary:Ryuichi's commits primarily focus on developing and refining a PyTorch-based text-to-speech synthesis model. The commits involve adding new features, experiment scripts and adjusting model configurations (including layers and kernel sizes). They appear to be actively working on the architecture of the DeepVoice3 model, the integration of attention mechanisms, and implementing techniques to improve the performance and robustness of the speech synthesis process.
pytorchspeech-to-textsynthesisconvolutionalneural-networks
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.