Computer Science & Engineering at Konkuk University · Seoul, Korea
Student developer building backend systems and exploring AI integrations. Always eager to learn, build, and keep my architecture clean.
Recent work includes:
- a financial backend that fixes a hypothetical position at purchase time and calculates its later opportunity cost,
- a KIPRIS → OpenCLIP → FAISS pipeline for trademark-image retrieval,
- and a distributed shared editor using Ricart–Agrawala mutual exclusion and Lamport clocks.
Java / Spring Boot · PostgreSQL · Python / FastAPI · Data-intensive Systems
Portfolio · English · 포트폴리오 · 한국어
| Project | What I worked on | Stack / current scope |
|---|---|---|
| IfSave 🔒 · Case Study | Built the backend/data/AI flow for comparing a purchase with 60 predefined assets (30 Korean and 30 US). A purchase fixes hypothetical quantities first; later prices are used for opportunity-cost calculations. Manual/CSV/OCR inputs and six AI personas are connected while numeric calculation stays outside the LLM path. | Java 21 · Spring Boot 3 · PostgreSQL 16 · private repository |
| Distributed Shared Editor | Implemented Ricart–Agrawala distributed mutual exclusion, Lamport-clock ordering, REQUEST/REPLY/HELD/RELEASE handling, WebSocket synchronization, and late-comer session sync. | Java 17 · Spring Boot · JPA · H2 · STOMP/SockJS · Demo |
| MarkLens | Worked on the trademark-search data path: KIPRIS collection/cleanup, image extraction, OpenCLIP 512D embeddings, FAISS Top-K retrieval, and FastAPI serving. The current develop track uses a limited 1,000-trademark research sample and reports visual similarity only. | Python 3.11 · FastAPI · OpenCLIP · FAISS · PostgreSQL |
| ReGrip | Connected four Canvas rehabilitation games to a replaceable DataService layer (localStorage ↔ REST), session/XP records, and offline-first persistence. |
FastAPI · SQLAlchemy · Vanilla JS · Canvas 2D · software/simulation validation |
2025.03–2025.08 · Android team project
- 16 pull requests merged into the original team repository under my GitHub account.
- Main areas: My Routine, Routine Feed, search, profile/follow flows, navigation, API integration, and FCM notifications.
- In the 2026 portfolio copy, I separately improved network boundaries, failure-state handling, tests, CI, and documentation.
- Backend — Java, Spring Boot, JPA, REST API, Python, FastAPI
- Data — PostgreSQL, SQLite, Flyway, FAISS, OpenCLIP
- Reliability — JUnit, pytest, Vitest, Playwright, GitHub Actions, lint/typecheck
- Client integration — TypeScript, Next.js, React, Kotlin, Jetpack Compose
- cv-panorama-stitching — OpenCV panorama-stitching pipeline built around a 10-million-pixel memory budget and regression tests.
- MyCo-Kit — React/TypeScript landing page for a circular-economy STEAM mycelium kit. Live
- Business Planner Skills — 15 reusable skills for Korean startup-support business plans using PSST (Problem, Solution, Scale-up, Team), with Claude Code and Codex installation guidance. English guide
2020.03–PresentB.S. in Computer Science & Engineering, Konkuk University2026IfSave Dream Semester · ReGrip Gyeonggi Youth Gap Year · MyCo-Kit U300 Growth Track
Project status and validation scope follow the corresponding repositories and project documents. Unverified production, clinical, or legal-performance claims are kept separate from implemented functionality.