I build backend systems, then stay around to see how they behave.

FastAPI and Spring Boot are my usual tools.
I enjoy the part after “it works”: deployment, load testing, failure analysis, and the next iteration.

_{만들고 끝내기보다, 직접 운영하고 측정하면서 다음 개선점을 찾습니다.}

01 / WHAT I BUILD

🚗 Streaming systems MQTT, Kafka, time-series storage, consumer design, and load tests that check the final stored result—not just queue health.

📄 Deterministic pipelines Parsers, layout engines, stable IDs, snapshot tests, and one-way data flows with explicit layer boundaries.

🔐 Reliable web services Authentication, payments, async workers, caching, monitoring, recovery paths, and security-conscious API design.

02 / SELECTED WORK

Telemetrix · Java Vehicle telemetry platform built around MQTT → Kafka → storage and anomaly consumers. load testing architecture decisions time-series data Found a silent 50% data-loss bug and recovered storage throughput to 98.5% of target.	Han-Flow · TypeScript Read-only macOS HWPX viewer with a one-way parser → layout → paint → PDF pipeline. parser boundaries deterministic IDs visual regression The parser never knows React; the layout engine never knows ZIP or XML.
AI Exam Generator · Python A deployed exam-generation service with payments, asynchronous workers, monitoring, and recovery paths. FastAPI Redis workers service operations Built the service, put it online, and stayed responsible for what happens when external APIs or workers fail.	Touch Bridge · Embedded / Python Accessible IoT interface that helps visually impaired users operate existing home appliances through voice commands. Flutter FastAPI Raspberry Pi / ESP32 Connected an accessible mobile interface to physical appliance control without requiring users to replace their existing devices.

More things I have built

• No-Click — AI-assisted review analysis for promotional content and missing drawbacks
• HoneyRest — a booking platform split across user, host, and admin applications
• AIDA — a training-data quality experiment using KITTI, YOLOv8, and 13 injected error conditions
• C Data Structures — lists, stacks, queues, AVL trees, graphs, hash tables, tries, and sorting algorithms implemented in C
• Linux C System Programming — files, processes, memory, signals, sockets, pthreads, and IPC

03 / TOOLBOX

Languages & backend

_{C foundations: pointers and dynamic memory, data structures, and Linux system programming with processes, signals, sockets, pthreads, and IPC.}

Data & messaging

Build, ship & observe

04 / GITHUB IN MOTION

_{How the language card is calculated}

_{honeyRest_host is excluded because it vendors about 11.4 MB of third-party JavaScript libraries under its static assets. The chart weights both code size and repository count; it shows repository composition, not a proficiency score.}

05 / RIGHT NOW

studying       Information Security @ University of Suwon
measuring      telemetry throughput, stored-data integrity, and query bottlenecks
building       a deterministic HWPX rendering pipeline
learning       where backend design meets security, IoT, and product operations

Good systems are not just built. They are observed.

If you want to talk about backend systems, strange bugs, or an ambitious project: