강형욱 (Hyeonguk Kang)
Music & Audio AI Researcher (aspiring) · Software Developer | 25 Years Old
4th year Information Security, Chosun University
Graduated Summa Cum Laude
Hello, I'm Kang Hyun-guk, who wants to study Music AI.
Based on my musical performance experience and absolute pitch, I became interested in how humans perceive and understand music — and this curiosity led me to explore AI models that allow machines to understand and generate music and audio.
Currently, I am preparing for audio/music generation and analysis research with the goal of an integrated music producer AI agent. My goal is to build an AI system that can deeply understand the structure and expression of music, and ultimately generate new music.
Research Interests
MIR
Music Information Retrieval algorithms and precise feature extraction from raw audio.
Representation
Self-supervised learning for robust audio representations in latent space.
ASR
Acoustic Scene Recognition for contextual environmental sound understanding.
Generative Music
AI-driven composition and expressive performance synthesis using LLMs.
Musical backgruond
Silver Prize for the National Seogyeong Music Competition
2010 / 11 | piano
Featured Projects
Planet Scope Data Analysis
Achieved 93.5% classification accuracy on satellite data using advanced Random Forest ensembles. Processed over 21,267 unique data points for environmental monitoring.
MOVE: IoT Location Tracking
Real-time indoor positioning system using ESP32 microcontrollers, MQTT protocols, and WebSockets for low-latency data transmission. Successfully signed MOU with university for deployment.
KTAS Prediction System
AI Challenge winner. Developed an intelligent prediction system using RAG (Retrieval-Augmented Generation), LLMs, and Vector Databases for highly accurate triage assessments.
Bitcoin Trading
LSTM-based predictive modeling for cryptocurrency market trends.
AI Travel Agent
AWS-cloud architecture for personalized itinerary recommendations.
Gemini Scheduler
Schedule optimization using Google Gemini API for productivity.
Figma-MCP Diary
Integrated diary application utilizing Model Context Protocol.
Studied Papers
- VirtuosoNet: A Hierarchical RNN-Based System for Modeling Expressive Piano Performance
- CP-Mobile: Distilling the Knowledge of Transformers and CNNs
- Heterogeneous Sound Classification with the Broad Sound Taxonomy and Dataset
- Deep Residual Learning (ResNet) for Computer Vision and Audio Recognition
- LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation
Honors & Awards
NASA Space Apps Challenge 2025
2nd Place
AI Challenge Season 2
Grand Prize (최우수상)
Chosun University Info Security
Top Student (Summa Cum Laude)
Metamobility Autonomous Driving
Grand Prize (최우수상)
SW Centric Univ. Open Source Idea
Gold Prize
LLM Hackathon
Excellence Prize
SW Centric Univ. Excellence SW Membership
Grand Prize (대상)
National Music Competition
Silver Medal (2010.11, Piano)