Backend Software Engineer | Distributed Systems | Java/Spring | AWS/EKS

I build backend systems where correctness, scale, and operating cost matter. At Samsung Electronics, I work on high-traffic production platforms with a focus on database migration, IAM modernization, batch processing, and cloud-native infrastructure.

LinkedIn | GitHub | Blog | Email

• Distributed backend systems with Java, Kotlin, Spring Boot, and production-grade data stores
• Large-scale database modernization across DynamoDB, Cassandra, and relational systems
• Cloud-native infrastructure on AWS, EKS, Kubernetes, Docker, Jenkins, and GitHub Actions
• Reliability-focused engineering: zero-downtime migration, data integrity, cost reduction, and incident recovery
• Agent-based engineering harnesses for production-code delivery, including branch hygiene, verification gates, review loops, and PR handoff

I am also interested in agent-based harness engineering: turning AI-agent workflows into repeatable delivery systems with scoped commits, validation gates, fresh review loops, and clean PR handoff. I apply this approach in day-to-day production backend work, and I also use it to identify recurring personal workflows worth automating. ai-news-alerts is one example: a daily Slack brief built with my Agent Harness Playbook and local Codex Config skills/hooks.

Public repositories that show adjacent tooling and delivery interests.

Public, sanitized playbook for agent-based engineering harnesses: branch preflight, plan-first execution, dry-run safety, verification gates, fresh reviewer loops, and PR handoff. Together with Codex Config, it supported ai-news-alerts, a safe scheduled workflow for a recurring personal need.

Agent Workflows GitHub Actions Dry Run Safety Review Gates Operational Automation

Public allowlisted backup of my local Codex setup, including reusable skills, hooks, and safety guardrails. It shows the executable side of the harness: local workflow instructions, pre-tool safety checks, sync automation, and project-specific skills used alongside the playbook.

Codex Skills Hooks Automation Guardrails Workflow Safety

MCP-based document indexing and search server for personal Notion, Github, and Tistory blog content, powered by LlamaIndex, ChromaDB, and a custom tool API.

Python LlamaIndex ChromaDB FastMCP Notion API Asyncio

React and Flask image application built to practice end-to-end CI/CD, container delivery, and Kubernetes deployment workflows.

React Flask Docker Kubernetes Helm Jenkins GHCR

Learning archive application with categorization features and a Kotlin/Spring backend.

React Kotlin Spring Boot MongoDB

Awarded Grand Prize (1st place) in the Apache Zeppelin Open Source Contribution Program for impactful upstream contributions across code quality, reliability, and maintainability.

• ZEPPELIN-6306 Prevented NPE in StaticRepl with fail-fast JavaCompiler validation.
• ZEPPELIN-6299 Refactored StaticRepl for readability, correctness, and modern Java usage.
• ZEPPELIN-6264 Refactored InfluxDBInterpreter for maintainability.
• ZEPPELIN-6242 Improved Neo4jConnectionManager enum parsing and default handling.
• ZEPPELIN-6220, ZEPPELIN-6243, ZEPPELIN-6285, ZEPPELIN-6300 Additional cleanup and refactoring contributions merged upstream.

• Kubernetes/website#52238 Synced Korean ingress-minikube documentation with the English version.