Jaeyeon Kim
Machine Learning Engineer at Toss Securities
Seoul, South Korea
Summary
🤩
Rockstar🎓
Top SchoolJaeyeon Kim is a Machine Learning Engineer with 8 years of experience specializing in MLOps and ML platform engineering, particularly building Kubernetes-native AI infrastructure. He has a track record of accelerating ML delivery—leading a Kubeflow/MLflow-based platform that cut project lead time by 70% and re-architecting an inference proxy in Go that halved P99 latency and reduced resource use by 90%. At MakinaRocks he optimized GPU serving (reducing cold-starts by 30%) and designed multi-format model serving for scalable LLM deployments, and he continues to influence upstream projects as a Kubeflow/Katib reviewer and contributor. His background in mathematics (KAIST) and experience across backend, DevOps, and ML research give him a rare blend of theoretical rigor and production-grade engineering. Now based in Seoul and working at Toss Securities, he remains focused on developer-centric MLOps ecosystems and open-source collaboration.
8 years of coding experience
6 years of employment as a software developer
Bachelor's degree Physics (major) Mathematics (minor), Bachelor's degree Physics (major) Mathematics (minor) at Korea UniversityMaster's degree Mathematics, Master's degree Mathematics at Korea Advanced Institute of Science and Technology
Github Skills (20)
kubernetes10
python10
testing10
go10
kubernetes-pods10
hyperparameter-tuning10
automated-machine-learning10
postgresql9
command-line-interface9
dockers9
docker9
devops9
command-line9
sdk9
cli9
Programming languages (15)
Github contributions (5)
kubeflow/katib
Jun 2021 - Oct 2022
Automated Machine Learning on Kubernetes
Role in this project:
Back-end Developer & DevOps Engineer
Back-end Developer & DevOps EngineerContributions:144 reviews, 9 commits, 9 PRs in 1 year 4 months
Contributions summary:Jaeyeon made several contributions related to improving the Katib platform's backend and infrastructure. They implemented and validated Bayesian optimization algorithm settings, and also modified the SDK, which involved refactoring code and changing the return type of list APIs. Furthermore, the user addressed several bug fixes related to test timeouts, filter by experiment names and changing default metrics collect format. Additionally, they added health check endpoints and implemented support for PostgreSQL as a database backend for Katib.
xgboosthyperparametermachine-learninglightgbmautoml
kubernetes/minikube
Feb 2020 - Mar 2021
Run Kubernetes locally
Role in this project:Backend & DevOps Engineer
Backend & DevOps EngineerContributions:6 reviews, 12 commits, 12 PRs in 1 year 1 month
Contributions summary:Jaeyeon primarily contributed to bug fixes and improvements related to the `minikube delete` command, ensuring proper cleanup of Kubernetes context and configuration files. They updated tests to improve assertion logic and debugging. Furthermore, the user implemented changes to allow setting the kubeconfig path via environment variables and refactored the `cp` command to enhance usability and correct behavior, especially on Windows.
containerslocallyk8skubernetescncf
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.