Daehyun Kang

Education

Hankuk University of Foreign Studies B.S. in Computer Engineering, College of Engineering Mar 2023 - Present

Hanyang Science and Technology High School Diploma in Computer Networking Mar 2020 - Jan 2023

Experience

Vitasoft Sep 2021 - Mar 2023
AI Team · Software Engineer

Fellowships

Best of the Best Jul 2025 - Feb 2026
Next-Generation Security Leader Program · Security Product Development Track

Whitehat School Apr 2024 - Sep 2024
Next-Generation Security Leader Program

Software Maestro Apr 2023 - Dec 2023
Advanced Software Talent Development Program · Cohort Representative

Projects

HUFS CODE MASTER - Algorithm Learning Platform for HUFS Apr 2026 - Present
Team project | Live production service

• Backend

  • Built an online judge powered by DOMjudge judgehosts
    • Supported C, C++, Java, and Python submissions with exact, whitespace-insensitive, floating-point, and custom judging
    • Delivered real-time judging progress and per-test-case results through Server-Sent Events
  • Implemented authentication and user management with HUFS SSO
    • Decoupled SSO user UUIDs from internal service IDs and applied Redis-backed session authentication
    • Built role-based access control and audit logging for students, instructors, professors, and administrators
  • Implemented learning history and motivation features
    • Provided weekly solve counts, current and longest streaks, activity heatmaps, tiers, and per-problem progress
    • Designed answer processing to update learning history and track progress consistently in a single flow
  • Expanded the service into an always-on programming learning platform with classrooms, exams, quizzes, a community, and administrative tools

• AI

  • Built a local LLM-powered code feedback system
    • Structured problem statements, submitted code, test cases, expected and actual output, errors, execution time, and memory usage as model input
    • Provided root-cause analysis and improvement guidance along with time-complexity analysis and relevant algorithm concepts
    • Decoupled the Spring Boot backend from the AI inference server through an asynchronous job and callback workflow
    • Built an operational workflow with request-state tracking, automatic cancellation of long-running jobs, and result notifications
    • Adopted an Ollama-based local model to avoid recurring commercial API costs, with request queues and usage limits for constrained hardware

• Infrastructure

  • Built CI/CD pipelines with GitHub Actions
    • Ran Gradle builds and tests for backend pull requests, then built and transferred JAR files to a Mac mini server and restarted the Spring container automatically for production deployments
    • Automated frontend TypeScript checks and production deployments to Vercel
  • Deployed Spring Boot, MariaDB, and Redis on a Mac mini with Docker Compose
  • Configured a native Nginx reverse proxy and automatic TLS renewal, restricting application and database management ports to the local network

• Operations

  • Operated four learning tracks ranging from introductory to advanced
  • Ran a winter algorithm camp using learning materials developed by Professor Chan-Soo Shin
  • Continued operating and improving the service after its release as an official extracurricular program, using usage data and feedback from students and faculty
  • Worked with the Department of Computer Engineering and the Office of Information Technology to support programming courses and HUFS CODE FESTIVAL as an official learning platform

• Achievement: Grand Prize, 13th Bring Your Own Idea Competition (HUFS President's Award)

대현.com - Utility Service for Mafia42 Players Sep 2020 - Present
Personal project | Live production service

• Infrastructure
  • Built a CI/CD pipeline with GitHub Actions
    • CI: Automated Gradle builds, Docker image creation, and pushes to GHCR
    • CD: Automated GHCR pulls, blue-green container startup, and Nginx traffic switching on the production server
  • Monitoring
    • Built server and container monitoring with Prometheus and Grafana
• Backend
  • Improved performance for high-volume query APIs
    • Replaced request-time full-data computation with precomputed snapshot-based reads
    • Added composite indexes tailored to query patterns to improve critical database query performance
    • Introduced pagination and server-side search to reduce API response cost and frontend rendering load
      • Improved production database query time by approximately 7x and reduced API response payload by approximately 33x
• AI
  • Built an LLM-powered gameplay feedback system
    • Structured replay chat logs and game context to evaluate player decisions, communication, voting patterns, and team contribution
    • Designed a typed JSON response schema for scores, grades, team contribution, evidence, and improvement suggestions, turning AI output into service content
    • Decoupled the Spring Boot backend from a FastAPI-based LLM inference server through an asynchronous analysis pipeline
    • Built an operational AI workflow with analysis-state tracking, result persistence, automated feedback comments, and an administrator retry API
    • Connected the production server and GPU inference server over a private virtual network to avoid exposing the inference server publicly
• Operations & Monetization
  • Operated the production service continuously since 2020 and improved features based on user feedback
  • Built an advertising revenue model with Google AdSense and an in-house advertisement delivery and management system
• Achievement: 2,500+ registered users, 4,400+ MAU, and break-even reached

Control Tower - Integrated Cloud Monitoring Based on Log Intelligence ^GitHub Sep 2025 - Dec 2025
Best of the Best 14th team project | Open source

• Built an investigation-oriented monitoring system that correlates asset-change history and user activity logs by actor to provide the cause and context behind anomalies
• Implemented snapshot-based asset-change tracking and 30 declarative YAML policies
  • Mapped to AWS FSBP control IDs
• Calculated user-specific anomaly scores using Mahalanobis distance from baseline and decomposed findings into contributing features for explainable analysis
• Generated validation data by configuring an AWS environment and running a three-stage attack scenario: reconnaissance, permission testing, and privilege escalation
• Achievement: Published a paper at the CISC-W'25 Conference of the Korea Institute of Information Security & Cryptology

Grape Office - Username OSINT Tool for Korean Platforms ^GitHub May 2024 - Jul 2025
Whitehat School team project | Open source

• Designed a username-based OSINT tool optimized for Korean online communities
• Defined the user-identification data schema and implemented validation logic to reduce false positives
• Achievement: Released an open-source username OSINT project specialized for Korean communities

Collectime - Group Scheduling Service ^GitHub May 2023 - Nov 2023
Software Maestro team project | Live production service

• Designed and implemented RESTful APIs with Kotlin and the Spring Framework
• Built a CI/CD pipeline and automated Docker image builds and deployments
• Integrated external calendar services and built a batch synchronization system

Publications

A Behavior-Based Anomaly Detection Framework Using Cloud Service Logs
Korea Institute of Information Security & Cryptology CISC-W'25 Conference · First Author

Vulnerability reporting

CVE-2026-59336 Aug 2026
Spring AI · Low ^GitHub

CVE-2026-47883 Aug 2026
Spring Framework - Open Redirect in UrlHandlerFilter · Medium ^GitHub

CVE-2026-41729 Jun 2026
Spring Data REST - SpEL Injection via Map Key in JSON Patch · High ^GitHub

CVE-2026-40987 Jun 2026
Spring Integration - Remote-file Synchronizer Arbitrary File Write · High ^GitHub

Certifications

Google AI Professional Certificate Aug 2026
Google

TOEIC Speaking IM3 Jun 2026
TOEIC

2025 Daangn Builder's Camp Sep 2025
Karrot

Oracle Cloud Infrastructure Foundations Associate Jul 2025
Oracle

TOPCIT Level 3 Oct 2023
Institute of Information & Communications Technology Planning & Evaluation

Industrial Engineer Information Processing Sep 2022
Human Resources Development Service of Korea

Awards

Grand Prize, 13th Bring Your Own Idea Competition (HUFS President's Award) Jul 2026
Hankuk University of Foreign Studies

Outstanding Project Award, Spring 2026 Capstone Design Jun 17, 2026
Department of Computer Engineering, Hankuk University of Foreign Studies

Outstanding Project Award, 1st G-RISE Capstone Design Competition Jun 2026
HUFS G-RISE Project Group

Grand Prize, 3rd HUFSummer Hackathon (HUFS President's Award) Jun 2024
Hankuk University of Foreign Studies

Excellence Award, 2nd HUFSummer Hackathon (Dean of Engineering Award) Jun 2023
Hankuk University of Foreign Studies

Community & Leadership

Google Developer Groups on Campus HUFS 2023 - 2026

• 7th Cohort Organizer
  • Led developer community operations and event planning
• 6th Cohort Core Member
  • Supported developer community operations
• 5th Cohort Core Member
  • Supported developer community operations

HUFS CODE FESTIVAL Organizing Committee 2023 - 2025
HUFS AI Education Center & Google Developer Groups on Campus HUFS

• Competition Director, 7th HUFS CODE FESTIVAL (Nov 2025)
  • Built a self-hosted, on-premise competition platform
  • Led end-to-end competition planning and operations
  • Managed the problem-setting and review process
• Problem Setter, 6th HUFS CODE FESTIVAL (Oct 2024)
  • Authored and reviewed competition problems
• Problem Setter, 5th HUFS CODE FESTIVAL (Nov 2023)
  • Authored and reviewed competition problems

Hanyang Science and Technology High School Assistant Instructor, Department of Computer Networking Jul 2023

• Mentored student teams in refining project ideas and defining development direction for an industry-academia apprenticeship program
• Helped resolve technical issues in web and game programming implementations

HUFS Software Volunteer Corps Feb 2024 - Jan 2025