Jeong Jang-young

AI Engineer & Deep Learning Researcher
"Curiosity drives innovation. Through rigorous research and bold experimentation, I transform complex AI theories into practical solutions that shape the future."
akillness38@gmail.com akillness.github.io github.com/akillness

Research Philosophy

I am deeply passionate about developing practical applications that integrate Deep Learning, the core model of AI technology. My goal is to solve complex problems based on intellectual curiosity and create substantial value through innovative AI solutions that bridge the gap between theoretical advancement and real-world impact. AI Specialist

Academic Excellence

Sep 2024 - Present
Ph.D Student in Game Engineering
Hongik University
Aug 2014 - Aug 2016
Master of Science in Game Engineering
Hongik University
Mar 2007 - Feb 2014
Bachelor of Science in Game Engineering
Hongik University

Core Strengths

Intellectual Curiosity
Bold Innovation
Reliable Execution
Results Driven
01

Technical Expertise

Programming Languages

Python C/C++ C# Objective-C Dart JavaScript SQL

AI & Machine Learning

PyTorch TensorFlow LangChain LlamaIndex OpenCV Transformers PEFT/LoRA RAG

Development Ecosystem

Visual Studio VS Code PyQt Flutter Unity3D Docker Git

Infrastructure & Platforms

GCP AWS Linux ONNX Qdrant Ollama
02

Ph.D Research Projects

Doctoral Research Program
Jul 2024 – Present
VLM-based AutoQA - Context Aware Bug Reporting
Feb 2025 – Present
  • Research on error detection utilizing VLM performance models based on VideoGameQA Bench
  • Extension of image-based AutoQA Macro Tool bug detection capabilities
  • Commercial-grade advanced development and optimization
VLM Computer Vision AutoQA Context Analysis
Multi-Agentic AI Framework with Ollama, LangGraph & PraisonAI
Apr 2025 – May 2025
  • On-device Ollama server and multi-agent workflow process using praisonaiagens
  • Knowledge Graph-based Semantic Routing implementation
  • MCP (Model Context Protocol) and Function Calling (Crawling) Handler development
Ollama LangGraph Multi-Agent Knowledge Graph MCP
LLM-based Game Design Agent Development
Mar 2025
  • Development of Vive GameDesigner Agent functionality utilizing ChatGPT API
  • Score-based chunking using ColBERT embeddings with Qdrant VectorDB
  • Task Role operations through Task Planning (planning, system, balance, event, scenario)
  • Function Calling for document processing (PDF, Word, Excel, PowerPoint) and web crawling
ChatGPT API ColBERT Qdrant Task Planning
03
LLM-based RAG & Agentic AI Construction Project
Feb 2025 – Mar 2025
  • Implementation of resume analysis system using LlamaIndex framework
  • Development of coding & web search Agentic AI system based on OpenAI API GPT-4o mini
LlamaIndex RAG GPT-4o Agentic AI
SLM-based On-Device Chatbot System Construction Project
Jan 2025 – Feb 2025
  • Performance evaluation and testing of SLM models TinyLlama, Deepseek R1 distill utilizing LangGraph, LangChain, Zero/Few-shot prompting, and Chain of Thought
  • Model structure separation and server system construction for GPT Chatbot Web Service implementation
  • Socket and Web Socket Server Multi-Process Python code implementation
  • Web Client Dart code implementation using Flutter
  • PEFT code implementation for pre-training systems
  • Post-training system exploration and reinforcement learning Cartpole environment project implementation including IMPALA, A2C, A3C, ACER
TinyLlama Deepseek R1 LangChain PEFT Flutter Reinforcement Learning
Image-based System-Independent AutoQA Tool Development
Sep 2024 – Jan 2025
  • PyQt5-based GUI implementation with OpenCV-based Multi-Scale Template Matching
  • Multi-thread pool (GIL) application for speed optimization
  • Process Handler implementation using Pywinauto for macro functionality
  • Keras-OCR application for GUI Context extraction functionality
  • Fine-tuning DETR (Detection Transformer) testing for gameplay verification
  • Fine-tuning YOLOv7 application for gameplay video verification
PyQt5 OpenCV DETR YOLOv7 OCR
On-Device SLM Chatbot Optimization for Detective Game Runtime Enhancement
Jul 2024 – Aug 2024
  • PEFT LoRA training for Polyglot-kr-1.3B model
  • LLM Python server optimization through batch separation and structural design optimization
  • CPU inference model ONNX exporting optimization for Unity Application service
Polyglot-kr PEFT LoRA ONNX Unity
04

Professional Experience

NC SOFT
Mar 2018 – Jul 2024 (~6 years 4 months)
Digital Double Facial 4D Scan and 4D HMC Mesh Low-poly Process Construction
Jun 2022 – Dec 2023
  • Latest version update of Faceform (formerly R3DS) system pipeline for 4D Scan Mesh Low-poly processing
  • Construction of Faceform system pipeline for 4D HMC (Head Mount Camera) Mesh Low-poly processing
3D Scanning Faceform Computer Vision Digital Twin
Robot Arm Programming for Camera Calibration Automation
Mar 2022 – Jun 2022
  • RobotArm utilization for Camera Calibration Checker board photography
  • Serial communication implementation for 5-actuator Dynamixel Openmanipulator-x (Python, C#)
Robotics Dynamixel Camera Calibration Serial Communication
Digital Twin Facial 4D Scan Mesh Low-poly Process Construction and Post-processing
Oct 2021 – Mar 2022
  • R3DS (Russia 3D Scan) Solution pipeline system construction for 4D Scan Mesh Low-poly processing
  • High-volume processing system pipeline construction and post-processing result output for 4D Scan Mesh Low-poly
  • Python code optimization for UV-based topology registration/reordering
Digital Twin 3D Processing Pipeline Optimization
05
CA (Cellular Automata) based Match-3 Puzzle Game Difficulty Generator Web Service Development
Mar 2021 – Oct 2021
  • Active user testing purposes for MATCH-3 Level generator
  • Python Flask-based web client functionality providing individual and *.zip format download of trained GAN model results through parameter adjustment
Flask GAN Cellular Automata Web Service
CA (Cellular Automata) based Match-3 Puzzle Game Difficulty Generator Development
Oct 2019 – Mar 2021
  • Utilization of procedural content generation methods for Match-3 Level Generator in game planning
  • CA (Cellular Automata) functionality implementation and CA model-based Level Generator development
  • Performance improvement through CA Generative Combination proposal for more diverse patterns
  • Utilization of GAN (Generative Adversarial Networks) model training data
Cellular Automata GAN Procedural Generation Game AI
MCTS (Monte Carlo Tree Search) based Match-3 Puzzle Game Difficulty Evaluation Agent Development
Mar 2018 – Sep 2019
  • Agent development for simulation-based evaluation of Match-3 puzzle game difficulty
  • MCTS (Monte Carlo Tree Search) functionality implementation and application to game Agent logic
  • Acceleration and achievement of meaningful results in Level difficulty evaluation
MCTS Game AI Agent Development Simulation
06
COM2US
Oct 2016 – Feb 2018 (~1 years 6 months)
"Dungeon Delivery" In-Game Reinforcement Learning Agent Application Testing
Jul 2017 – Dec 2017
  • Exploration and research on applying AI technology to developed games
  • TensorFlow-based reinforcement learning model design and game application through socket communication
  • Gridmap design for game environment as learning environment and DQN (Deep Q-Network) application to Action prediction model
  • Client-Server Network configuration between Unity3D and Python platforms for environment information and Action execution, and learning progression
TensorFlow DQN Reinforcement Learning Unity3D
Developer Collaboration Site Issue Tracking Development for Process Knowledge Sharing
Apr 2017 – Jul 2017
  • In-house web service development providing Stack Overflow-like functionality for developer collaboration process sharing
  • Development using MVC (Model View Controller) Pattern with CodeIgniter (CI) framework, WYSIWYG editor SummerNote, DB (MySQL), and AJAX functionality
CodeIgniter MySQL MVC Pattern Web Development
Additional Projects and Responsibilities
Jan 2017 – Apr 2017 / Nov 2016 – Jan 2018
  • In-house website management (Front-End/Back-End/DB), 2 internal hackathon Unity3D game development projects
  • New hire/experienced hire recruitment site management and process assistance, mobile-related development/optimization technical R&D
  • Maintenance of discontinued mobile game services
Full-Stack Development Mobile Optimization R&D
07

Intellectual Property

Game Difficulty Determination Method and Apparatus
Aug 2019 – Dec 2020
Patent No. 10-2195-4710000
Emotion Computing Device and Its Operation Method
Nov 2017 – Dec 2017
Patent No. 10-2017-0009076

Professional Activities & Recognition

IEEE RAAI2024 Conference | Poster Presentation
Dec 20, 2024
  • "Image-Based Game Quality Assurance Automation Process"
IEEE Conference QA Automation
NC Research AVATAR Promotional Video
Apr 17, 2023
NCDP2020 Conference Presentation
Apr 28, 2020
  • "MATCH-3 Puzzle Game Difficulty Generator Development" Presentation
NCFellowship Program Management
Jan 2019 – Dec 2022
JTBC Entertainment Program Technology Transfer
Dec 2017
  • Patent technology transfer related to "The Theoretically Perfect Man"
"Love Sensor" HCI APP MARKET Encouragement Award
Jan 28, 2016
08

Publications & Research Output

HCI Korea 2016 Paper
Jan 28, 2016
  • Jeong Jang-young, Kim Young-bin, Lee Sang-hyuk, Kang Shin-jin, "Facial Expression Change Detection System for Game Play Analysis", HCI Korea, 2016
Computer Vision HCI Emotion Analysis
PeerJ Journal Publication (SCI-E)
Dec 10, 2015
  • Y. B. Kim, S. J. Kang, S. H. Lee, J. Y. Jeong, H. R. Kam, J. Lee, Y. S. Kim, J. S. Lee and C. H. Kim, "Efficiently Detecting Outlying Behavior in Video-Game Players", PeerJ, 2015 (SCI-E)
SCI-E Behavioral Analysis Game Analytics
EmotionTracer Emotion Analysis Platform G-star Exhibition
Nov 12 – Nov 15, 2015
  • G-star exhibition interview participation
Exhibition Emotion Analysis G-star

Key Achievements Summary

Research Excellence: Leading cutting-edge AI technology research including VLM, LLM, and Multi-Agent AI through Ph.D studies, developing practical AI solutions based on over 8 years of industry experience

Innovation Impact: Holder of 2 patents, published SCI-E papers, participated in IEEE conferences, combining academic achievements with commercialization experience through practical projects at NC SOFT and COM2US

Technical Leadership: Demonstrated responsible leadership and technical organization management capabilities through NCFellowship program management, leading various AI projects, and team mentoring

Future Vision: Continuously pursuing innovative research that transcends AI technology limitations based on intellectual curiosity and challenging spirit, serving as a bridge between industry and academia
09