Kim Minkyu
Machine Learning Engineer at Toss Securities
South Korea
Summary
👤
Senior🎓
Top SchoolKim Minkyu is a Machine Learning Engineer with nine years of experience building production ML systems, from ETL and recommender pipelines to NLP and personalized advertising at scale. Based in South Korea, he has driven end-to-end ML work at Toss Securities and NAVER, combining research-grade modeling with robust workflow and infra skills (Airflow/argo, Kubernetes, Docker, Hadoop ecosystem). He contributes to open-source ML tooling—evidenced by implementing and extending SpecAugment across TensorFlow and PyTorch—and focuses on practical, reproducible deployments. His background spans research (ETRI) to applied finance and e-commerce, giving him a strong sense for data quality, debiasing, and measurable model performance.
9 years of coding experience
10 years of employment as a software developer
Master of Science (MS), Computer Science, 3.79/4.3, Master of Science (MS), Computer Science, 3.79/4.3 at 서울대학교Bachelor of Science (BS), computer science, 3.48/4.3, Bachelor of Science (BS), computer science, 3.48/4.3 at 한국과학기술원(KAIST)
English
Github Skills (6)
data-augmentation10
python10
tensorflow9
speech-recognition9
pytorch9
machine-learning8
Programming languages (2)
Github contributions (5)
DemisEom/SpecAugment
Apr 2019 - Nov 2019
A Implementation of SpecAugment with Tensorflow & Pytorch, introduced by Google Brain
Role in this project:
ML Engineer
ML EngineerContributions:26 commits, 1 PR, 27 pushes in 7 months
Contributions summary:Kim primarily contributed to the implementation and improvement of SpecAugment, a data augmentation technique for speech processing. Their work involved modifying the `spec_augment.py` file by updating remarks, fixing typos, refactoring code, and integrating time warping using Tensorflow's `sparse_image_warp`. Furthermore, the user added a PyTorch version and test code for the SpecAugment algorithm. This indicates a focus on enhancing and expanding the functionality of the SpecAugment implementation across different frameworks.
pytorchbrainpythondeep-learningspeech-recognition
DemisEom/RNNT-pytorch
May 2019 - Jun 2019
Contributions:25 commits in 1 month
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.