Backend & AI Developer based in Seoul, South Korea.
Passionate about backend developments, and server systems with a strong interest in AI. Always eager to learn new technologies and ready to adapt, contribute, and communicate effectively in any environment.
🔗 LinkedIn | 🐈⬛ GitHub | 📝 Tistory
Dongguk University · Seoul, Korea · Mar 2021 – Feb 2027 (Expected)
Bachelor of Engineering in Computer Science and Engineering
- GPA: 4.06 / 4.5
- Academic Excellence Scholarship: Spring 2021, Fall 2021, Spring 2022, Fall 2022, Fall 2025, Spring 2026
Backend & ML Developer | Mar. 2026 – Present
Automated Complaint Classification & Department Dispatch System
- Distributed Processing: Implemented asynchronous inter-service communication using Apache Kafka across three topics — embedding trigger, batch file classification request, and single-complaint classification request — with callback-based result propagation.
- AI Pipeline: Applied a two-stage RAG pipeline — KURE-v1 embeddings + pgvector cosine similarity for process code assignment, then top-N retrieval + Qwen3.6 27B for department dispatching.
- Multi-Environment Configuration: Separated deployment configs across two environments — on-premises (Docker Compose) and AWS ECS — using Spring profile-based YAML files, environment-specific Dockerfiles, and per-environment Nginx configurations.
- LLM Integration: Self-hosted Qwen3.6 27B on a dedicated server using Ollama and integrated it via the OpenAI-compatible REST API, enabling seamless replacement without SDK changes.
- CI/CD: Automated the full deployment pipeline with GitHub Actions — on push to main, builds a Gradle JAR, packages three Docker images (Spring Boot, FastAPI, Nginx), pushes them to AWS ECR with commit SHA tags, and force-deploys each service to AWS ECS by registering updated task definitions with injected GitHub Secrets.
Backend & ML Developer | Mar. 2026 – June. 2026
Procurement Demand Forecasting Service Based on Weather Data
- Modeling: Trained two LightGBM regression models to predict the number of public procurement announcements based on weather, date, and auto regressive-lag features.
- Async Event Pipeline: Designed a Kafka-based async pipeline — Spring Boot scheduler triggers daily data ingestion (G2B, KMA, Holiday APIs), publishes events to Kafka, and the ai-server consumes them to run classification and forecasting sequentially.
- Backend System: Built REST APIs with Spring Boot and FastAPI; real-time prediction results delivered to the frontend via Server-Sent Events (SSE). PostgreSQL with pgvector for vector similarity search, Flyway for schema migration, Nginx as reverse proxy.
- Deployment: Containerized the entire backend service with Docker Compose and visualized results through a dashboard interface.
Backend Developer | Jan. 2026 - Feb. 2026
Remote Smishing Link Execution & AI-Based Threat Analysis Service
- Containerized Execution: Built and optimized Docker images to safely execute suspicious links in isolated remote environments.
- Threat Scoring: Designed a malicious scoring system to quantify risk levels based on execution results and heuristics.
- Cloud Architecture: Leveraged AWS Fargate to dynamically spawn analysis containers, isolating potentially harmful workloads from the main system.
- AI Explanation: Integrated AI-based analysis to provide interpretable summaries of detected threats.
Backend Developer (ETL & Data Pipeline) | Sep. 2025 - Dec. 2025
Conversational Mobile App for Location-Based Discount Recommendations
- Data Pipeline: Designed and implemented an ETL pipeline to collect and normalize promotional data from financial institutions and partner services.
- Web Crawling: Built a scheduled crawling system using Playwright, automating data extraction and ingestion into the database.
- AI Integration: Supported RAG (Retrieval-Augmented Generation) architecture by managing vector embeddings and data preprocessing for recommendation quality.
- Data Reliability: Ensured data freshness and consistency through periodic batch processing and validation logic.
- Dongguk University Computer Engineering Student Council - Administrative Staff | 2021
- Dongguk University Computer Engineering Student Council - Head of Administration | 2022
- GDG (Google Developer Groups) - Study Participant | Winter 2025
- ORACLE - Seminar Participant | Mar. 2026
- UMC (University MakeUs Challenge) - Challenger (10th) & Server Part | Mar. 2026 – Present
- 📧 tdw0312@gmail.com
- 📱 +82 10 8767 4204