DevOps Portfolio

Kubernetes · Autoscaling
CI/CD · Test Automation
AIOps · Developer Platform
Cloud Research

Hello, I’m Donggyun Kim

This is my

Portfolio.

김동균 Donggyun Kim

DevOps · Cloud Engineer

Connecting where products begin
to how they run.

I have researched ways to optimize resource utilization and response latency in cloud environments.
At Reco, I chose a PO/QA role to understand the product as a whole and took on product planning and release QA.
I reduced repetitive manual work with E2E automation and a RAG platform, and continue to build and improve infrastructure to increase productivity across the engineering organization.

I have optimized performance and resource use in research, and QA time and quality in practice.
My strength is analyzing a problem and proactively improving it to raise work efficiency.

Experiences.

2025.03

Reco | Product Owner · QA

From requirements to release verification: QA automation, RAG platform, internal infra
2024.03

Korea University | Research Mentor · TA

Mentored 4 undergrad researchers; 4 papers (1 silver award) · TA for 5 courses
2023.03

Korea University | Graduate Researcher

Distributed & Cloud Computing Lab · Kubernetes autoscaling research, 12 papers
2021.12

COU | Founder · PM

Open-source online judge, grown from a 3-person project into a 10-person startup

Education.

Master's, CS & Engineering

Korea University

2023.03 – 2025.02

GPA 4.11 / 4.5

Bachelor's, Computer Engineering

Hallym University

2016.03 – 2023.02

GPA 3.74 / 4.5

Tools.

Kubernetes · Istio · Docker

GitHub Actions · Playwright

Go · Python · TypeScript

Prometheus · Grafana

Focus.

Autoscaling

CI/CD · QA Automation

AIOps · RAG

On-prem Provisioning

Awards.

ICSOC 2024 Distinguished Paper

Hallym University Graduation Merit Award

2022 SW Talent Festival, Excellent Work

SW Hackathon, 3rd Place ×2

Dean's List ×3, Best Semester Performance ×1

INDEX.

01

RAG Knowledge Base

Internal knowledge search connecting Jira, Confluence, and Git

AIOPS · PLATFORM
02

E2E QA Automation

E2E and k3s verification infrastructure built from manual QA

TEST AUTOMATION · CI/CD
03

LARE-HPA

Kubernetes autoscaler co-optimizing latency and resource efficiency

RESEARCH · ICSOC 2024
04

Product Ownership

Re-contract Redesign · Bulk Copy · Scheduled Suspension

PRODUCT · PO
05

More Works

k8s-cluster-bootstrap · KUCODE · COU

AUTOMATION · OPEN SOURCE

Notable Development Work 01 · AIOps

RAG Knowledge Base

View Project

I built an MCP-based Agentic RAG platform that lets agents retrieve policies and code distributed across Jira, Confluence, and Git. It reduces the need to find and supply context for each task, supporting QA preparation, development assistance, and policy review of planning tickets.

Role: Solely responsible for designing, developing, and operating the internal RAG platform (retrieval, MCP server, policy review, monitoring)
Stack: Go · Python · BGE-M3 · Qdrant · MCP · Claude Code · Prometheus · Grafana
Outcome: Internal knowledge search platform built · adopted for QA and development · Jira policy reviews automated

+6.7 %

implementation correctness with RAG

+13 %

more tickets verified per day · vs. E2E testing alone

-26 %

post-deploy bug rate reduction · vs. E2E testing alone

RAG Platform and Jira Policy QA Architecture

Notable Development Work 02 · Test Automation

E2E QA Automation

View Project

Connecting QA records through Google Sheets and Jira still left basic checks to people. I designed and built E2E infrastructure from zero to address that limit and connected Admin–Driver workflows. I then built isolated k3s environments to verify basic functionality in PRs before deployment to the test server.

Role: Solely responsible for building and operating QA automation (test scenario design, CI, and per-PR k3s environment setup)
Stack: TypeScript · Playwright · GitHub Actions · k3s · Helm
Outcome: Company's first E2E automation infrastructure built · per-PR pre-deployment verification introduced

88.2 %

test coverage

+31 %

more tickets verified per day

-44 %

post-deploy bug rate reduction

Notable Development Work 03 · Research

LARE-HPA

View Project

A Kubernetes autoscaler that retunes its threshold and cool-down on its own, fully online with no pre-training — co-optimizing the trade-off between latency and resource efficiency.

Role: First Author (method design and implementation, experiments)
Stack: Python · Kubernetes · Istio · Prometheus
Outcome: ICSOC 2024 Distinguished Paper Award · patent application filed · software copyright registered

50.34 %↓

Mean latency vs default HPA

97.24 %

SLO satisfaction

Top 6

ICSOC 2024 Distinguished

Product Ownership 04

Product, from the start.

I started in a PO/QA role to become a developer who understands a product from start to finish, not one who only writes code.
In the three projects below, I drew on my development background to lead the work from requirements to edge cases, designing policies that automate recurring operations.

Re-contract Process Redesign

Re-contract process that automated the manual task-amount transfer and rescheduling to cut omissions.

Policy Matrix
Transfer Automation

Bulk Copy

Ops automation that turned one-by-one re-contracts into contract-based bulk copy, cutting repetition.

Async Batch
Whitelist Validation

Scheduled Suspension

Ops automation that moved same-day manual suspensions to future-dated scheduling to cut timing misses.

State Transitions
Proration Rules

More Works 05

Skills

Tech Stack.

Programming Languages

Proficient In

Python
Java
Go
TypeScript

Familiar With

C
C++
Bash

DevOps & Frameworks

Proficient In

Kubernetes
Docker
containerd
Prometheus
Grafana
GitHub Actions
Playwright
Linux
FastAPI
Django
Vue.js

Familiar With

Flask
PostgreSQL
Redis
PyTorch
Istio
Calico
Svelte

Others

Proficient In

Codex
Claude Code
Cursor
MCP
On-premise Cluster Building
AWS
GCP
Tencent
Unix/Linux

Familiar With

Qdrant
RAG
Figma

Research

Research Projects.

Government-funded R&D projects I contributed to, from undergrad and grad school through today.

01
Develop a smart waste work platform(Saas) via cloud
Research Personnel (leading day-to-day execution) · industry–academia collaboration
- R&D
- OCR
- AI
- Performance (TPS)
- Waste Management
02
Platform built to support open-source software education and promote open-source activities under the National Program of Excellence in Software
Research Personnel · Project Manager
- Open Source
- Full-stack
- Project Manager
03
Development of a smart aging public service platform and prevention of forest disasters using a forest information convergence cloud system
Research Personnel
- Edge Cloud
- Resource Management
- Autoscaling
- Real-time Processing
- High Availability
04
Development of an edge cloud-based vehicle sharing platform that supports user-specific automotive healthcare services
Research Personnel
- Car Sharing Platform
- Edge Cloud Computing
- Automotive Healthcare
- Driving Assistance Service
- Convergence Service Platform

Publications

Papers.

12: papers, intl. + domestic
3: SCIE-compatible top-tier
1: ICSOC Distinguished (Top 6)

International Conferences

In Preparation 2026
ARTT-HPA: Heterogeneity-Aware Autoscaling with Adaptive Resilient Thresholding and Traffic Allocation for Multi-Replica Microservices on Kubernetes (tentative title)

Donggyun Kim, Hokun Park, Heonchang Yu
- On-going
- Co-Author
IEEE CLOUD 2025 2025
HEART: Heterogeneous-Aware Traffic Allocation in Multi-Replica Deployments on Kubernetes

Hokun Park, Donggyun Kim, Hyungjun Kim, Gyujeong Lim, Heonchang Yu
- SCIE-compatible top-tier
Paper Link
ICSOC 2024 2024
LARE-HPA: Co-optimizing Latency and Resource Efficiency for Horizontal Pod Autoscaling in Kubernetes

Donggyun Kim, Hyungjun Kim, Eunyoung Lee, Heonchang Yu
- Distinguished Paper (Top 6)
- SCIE-compatible top-tier
Paper Link View Project
ICSOC 2024 2024
Optimizing Traffic Allocation for Multi-replica Microservice Deployments in Edge Cloud

Hokun Park, Hyungjun Kim, Donggyun Kim, Gyujeong Lim, Heonchang Yu
- SCIE-compatible top-tier
Paper Link
IC 2024 2024

Proactive Auto Scaling Based on Marginal Request Change Analysis for Reducing Tail Latency in Kubernetes Cluster

Donggyun Kim, Heonchang Yu

Paper Link

Domestic Conferences

KCC 2025

Dynamic Replica Replacement Technique Based on Deep Reinforcement Learning for Cost-Effective Cluster Utilization in Heterogeneous Kubernetes Environment

Seoeun Kwon, Minjae Kang, Donggyun Kim, Heonchang Yu

Paper Link
KIPS ACK 2024

An Analysis of CPU Utilization and Service Latency Based on Cool Down Time Configuration in Kubernetes

Seoeun Kwon, Donggyun Kim, Heonchang Yu
Silver Award
Paper Link
KIPS ACK 2024

An Analysis of Service Latency Based on Replica Resource Granularity in Kubernetes Environment

Geonwoo Kim, Donggyun Kim, Heonchang Yu

Paper Link
KIPS ACK 2024

An Analysis of Response Time Depending on CPU Resource Contention between Pods in Kubernetes Environment

Seungjoo Han, Donggyun Kim, Heonchang Yu

Paper Link
KCC 2024

Cost-Efficient Resource Allocation Technique using Bayesian Optimization in Serverless Environment

Minjae Kang, Donggyun Kim, Heonchang Yu

Paper Link
KIPS ACK 2023

A Performance Analysis of Pulling Rate Based on Image Layer Concurrency Level in Docker Environment for Cold Start Mitigation

Minwoo Kang, Donggyun Kim, Heonchang Yu, Jihun Kang

Paper Link
KCC 2023

A Performance Analysis on Auto Scaling Using Different Image Pulling Methods in Kubernetes

Donggyun Kim, Chunggeon Song, Hyungjun Kim, Heonchang Yu, JoonMin Gil

Paper Link
KIPS ACK 2022

Design and Implementation of Coding Platform Using Open-Source Software

Donggyun Kim, Jonggun Park, Jiho Kim, Doyeon Kang, Ivanova Ekaterina, Youngwoong Ko

Paper Link

Published while mentoring an undergraduate researcher

Awards

Awards and Honors.

13 awards & honors

2024

2024.12

Distinguished Paper Award (Top 6)

LARE-HPA

ICSOC 2024

2023

2023.02

Merit Award

Graduation Award

Hallym University

2022

2022.12

SW Scholarship (Competitions and Contests)

SW Festival & Workshop participation

Hallym University
2022.12

SW Scholarship (On-campus)

Excellence in SW program participation

Hallym University
2022.12

Excellent Work

2022 SW Festival

National Program of Excellence in Software
2022.10

3rd Place

SW Open-Source Programming Challenge Hackathon

Hallym University
2022.07

Dean's List

Academic Excellence, 1st Semester

Hallym University

2021

2021.07

Dean's List

Academic Excellence, 2nd Semester

Hallym University

2019

2019.12

Dean's List and Academic Excellence Scholarship

Academic Excellence, 2nd Semester

Hallym University
2019.12

SW Scholarship (On-campus)

Excellence in SW program participation

Hallym University
2019.11

3rd Place

SW Open-Source Programming Challenge Hackathon

Hallym University
2019.07

Ilsong Scholarship

Best Academic Performance, 1st Semester

Hallym University
2019.07

Dean's List and Academic Excellence Scholarship

Best Academic Performance, 1st Semester

Hallym University

Writing

Stranger's LAB.

A personal tech blog on algorithms and CS fundamentals — consistent writing that has grown into serious traffic.

st-lab.tistory.com

total pageviews: 2.96M
monthly visitors (12-month avg.): 4.9K
subscribers: 735

Thank You

Let’s build
something.

I am looking for opportunities to build reliable services and better developer environments.

Donggyun Kim · DevOps · Cloud Engineer

TOP

Portfolio.

About Me

Connecting where products beginto how they run.

Experiences.

Education.

Tools.

Focus.

Awards.

INDEX.

RAG Knowledge Base

E2E QA Automation

LARE-HPA

Product, from the start.

Re-contract Process Redesign

Bulk Copy

Scheduled Suspension

Programming Languages

Proficient In

Familiar With

DevOps & Frameworks

Proficient In

Familiar With

Others

Proficient In

Familiar With

Develop a smart waste work platform(Saas) via cloud

Platform built to support open-source software education and promote open-source activities under the National Program of Excellence in Software

Development of a smart aging public service platform and prevention of forest disasters using a forest information convergence cloud system

Development of an edge cloud-based vehicle sharing platform that supports user-specific automotive healthcare services

International Conferences

ARTT-HPA: Heterogeneity-Aware Autoscaling with Adaptive Resilient Thresholding and Traffic Allocation for Multi-Replica Microservices on Kubernetes (tentative title)

HEART: Heterogeneous-Aware Traffic Allocation in Multi-Replica Deployments on Kubernetes

LARE-HPA: Co-optimizing Latency and Resource Efficiency for Horizontal Pod Autoscaling in Kubernetes

Optimizing Traffic Allocation for Multi-replica Microservice Deployments in Edge Cloud

Proactive Auto Scaling Based on Marginal Request Change Analysis for Reducing Tail Latency in Kubernetes Cluster

Domestic Conferences

† Dynamic Replica Replacement Technique Based on Deep Reinforcement Learning for Cost-Effective Cluster Utilization in Heterogeneous Kubernetes Environment

† An Analysis of CPU Utilization and Service Latency Based on Cool Down Time Configuration in Kubernetes

† An Analysis of Service Latency Based on Replica Resource Granularity in Kubernetes Environment

† An Analysis of Response Time Depending on CPU Resource Contention between Pods in Kubernetes Environment

Cost-Efficient Resource Allocation Technique using Bayesian Optimization in Serverless Environment

A Performance Analysis of Pulling Rate Based on Image Layer Concurrency Level in Docker Environment for Cold Start Mitigation

A Performance Analysis on Auto Scaling Using Different Image Pulling Methods in Kubernetes

Design and Implementation of Coding Platform Using Open-Source Software

2024

2023

2022

2021

2019

Stranger's LAB.

Let’s buildsomething.

Connecting where products begin
to how they run.

Dynamic Replica Replacement Technique Based on Deep Reinforcement Learning for Cost-Effective Cluster Utilization in Heterogeneous Kubernetes Environment

An Analysis of CPU Utilization and Service Latency Based on Cool Down Time Configuration in Kubernetes

An Analysis of Service Latency Based on Replica Resource Granularity in Kubernetes Environment

An Analysis of Response Time Depending on CPU Resource Contention between Pods in Kubernetes Environment

Let’s build
something.