Dabi Ahn
Lead AI Engineer at MarqVision
Seoul, South Korea
Summary
🤩
Rockstar🎓
Top SchoolDabi Ahn is a Machine Learning Engineer with 12 years of experience building and productionizing AI systems across startups and industry leaders in Seoul. He has led AI teams and shipped end-to-end ML features for VR concert production—covering alpha matting, denoising, super-resolution, frame interpolation and 3D reconstruction from stereoscopic footage. Previously at NVIDIA he optimized large-scale NLP and audio models and contributed to high-profile projects like FasterTransformer and CUDA-optimized FastSpeech, with hands-on CUDA and mixed-precision performance engineering. His research and open-source work on neural TTS and voice conversion (including the deep-voice-conversion repo) reflect deep expertise in speech synthesis, disentanglement of content and style, and low-data training. Comfortable moving between research prototyping and production engineering, he combines computer vision, audio, and systems-level optimization to deliver scalable ML pipelines. Outside work he’s a coffee enthusiast—an unexpected constant through long GPU training runs.
13 years of coding experience
10 years of employment as a software developer
Gyeongnam Science High SchoolTechnical University of DenmarkMaster of Science (MS) Computer Science, Master of Science (MS) Computer Science at Korea Advanced Institute of Science and Technology
English, Korean
Github Skills (11)
neural-network10
machine-learning10
eval10
voice-conversion10
deep-learning10
trainings10
tensorflow10
python10
evaluation10
modeling10
data-preprocessing9
Programming languages (3)
Github contributions (5)
andabi/deep-voice-conversion
Aug 2017 - Jul 2019
Deep neural networks for voice conversion (voice style transfer) in Tensorflow
Role in this project:
ML Engineer
ML EngineerContributions:106 commits, 3 PRs, 46 pushes in 1 year 11 months
Contributions summary:Dabi contributed to the development of deep neural networks for voice conversion. Their work involved refactoring code, updating evaluation procedures, and testing samples. The code changes focused on implementing model components and training/testing workflows. They were also involved in adjusting configurations.
deep-neural-networksstyle-transfertensorflowvoice-conversion
andabi/autoencoder-vae-gan
Apr 2017 - Jul 2017
Implementations of vanilla autoencoder, VAE, and GAN in Tensorflow
Contributions:40 commits, 27 pushes, 2 branches in 2 months
autoencodertensorflow
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.