콘텐츠로 바로가기
Hy's Blog
Bilingual Knowledge Archive
Hy's Blog
Public identityArchive-facing
hyunyoun's Blog

hyunyoun's Blog

Writing down the world I see and the self I build.

Identity surface

The public face of the archive stays inside the same system.

Projects, experience, and skills are presented as an editorial record rather than a detached portfolio template.

3 sections
Archive identityEN

Identity surface

About Me

Identity4 min read693 words

Overview of backend, platform, and DevOps experience, plus the technical areas I am focused on now.

Backend, Platform & DevOps Engineer

I've spent most of my career building backends and platform infrastructure with Python and Node.js. My work has ranged from public-sector B2G/B2B SaaS platforms to real-time high-availability notification systems and multi-cloud SRE monitoring — all across Azure, Naver Cloud Platform (NCP), and on-premise hybrid environments. I designed and operated these systems myself, and I got good at keeping them running.

Early on, shipping features was the whole job. At some point, though, the question shifted: "How fast does this system come back when something breaks?" That question stuck with me. Since then, I've been obsessed with HA architecture, observability, and building solid incident response workflows.

I've built CI/CD pipelines on NCP DevTools (Source Commit/Build/Deploy) and GitHub Actions, operated real-time anomaly detection with Prometheus, Grafana, and Alertmanager, and put together search (OpenSearch), caching (Redis), async processing (Celery), and ETL pipelines. The through-line has always been the same: availability and performance, at the same time.

When we migrated a public SaaS product to NCP-based microservices, I handled the CSAP, K-PaaS, and national security certifications directly. That meant addressing security requirements at the architecture level, not just checking boxes. On the performance side, I improved our notification pipeline to cut the error rate by 75% while scaling the system to handle 5x the load. Standardizing our monitoring brought MTTR down by 30%.

More recently, I've been expanding from my backend and infrastructure roots into data engineering pipelines, AI/LLMOps, and Infrastructure as Code.

What I'm focused on right now

I'm deepening the skills I already use in production while systematically learning new areas.

DevOps & GitOps

Topic Goal
ArgoCD + Helm GitOps workflows Implement canary and blue-green deployments hands-on
Terraform for NCP/Azure infrastructure Establish multi-cloud IaC patterns
GitHub Actions / NCP DevTools improvements Fully automate the test-build-deploy cycle
Kubernetes Operator patterns Operate custom resources via CRDs

Data engineering

Topic Goal
Apache Kafka + Schema Registry Build event-driven architectures from scratch
Apache Airflow ETL pipelines Design DAG-based scheduling and data workflows
Polars/Pandas data transformation High-performance data processing and analysis automation
dbt (Data Build Tool) Codify and test the transformation layer

Backend (going deeper)

Topic Goal
Go & Java/Spring Boot Expand backend capabilities beyond Python/Node
gRPC + Protocol Buffers Experiment with high-performance service-to-service communication
Async processing (Celery, Redis Queue) Push-test performance limits and optimize
Clean Architecture / DDD Apply domain-driven design in practice

Infrastructure & databases

Topic Goal
PostgreSQL HA (Patroni/pg_auto_failover) Battle-test failover in real operating conditions
Redis Cluster mode Develop recovery strategies for different failure scenarios
MinIO + Harbor Self-host private registries and object storage
InfluxDB / time-series databases Optimize time-series data management and queries

Observability

Topic Goal
OpenTelemetry-based distributed tracing Adopt distributed tracing as a standard
Loki + Tempo + Alloy stack Build a unified log/trace analysis pipeline
SLO/SLI definition & error budgets Quantify incident response effectiveness
eBPF-based observation Gain kernel-level performance visibility

Frontend (expanding toward full-stack)

Topic Goal
Next.js App Router + TypeScript Build production-quality UIs
React Query + Zustand Learn server state management patterns
API contract design Design backend APIs with UX in mind

AI / ML & LLMOps

Topic Goal
Ollama + LangChain + LangGraph Build local AI agents
MCP (Model Context Protocol) Tool integration and agent orchestration
PyTorch / Hugging Face Fine-tuning experiments and model serving basics
MLflow / Weights & Biases Experiment tracking and model management (MLOps)