Kim Jeong-won
M.S. Student, Computer Science and Engineering, Sungkyunkwan University · INGLab
Interests: AI Agents ・ Society ・ World, Runtime AI for Games
Projects
Team
Shiba Games : Practical research and business planning on LLM-based NPC Agents
Team Leader | 2024.06 ~ 2025.02
- A project for SPARK 14th
- Explored runtime AI for game NPCs using LLM APIs — from research to startup planning
ShibaRPG — RPG system with interactive NPC conversations and actions
2024.11 ~ 2025.02
- Built an RPG where players can converse, battle, and trade items with LLM-driven NPCs
- Validated user immersion through open storytelling
ShibaMarket — Interactive commodity trading game with LLM-based agents
2024.06 ~ 2024.10
- Designed and developed a trading game that utilizes LLM Agent's expressiveness and judgment for NPC interactions
RealBang : 3D room generation AI models and viewer based on Novel View Synthesis
Team Member | 2023.05 ~ 2024.01
- A project for SPARK 11st and Co-Deep Learning
- Compared the performance metrics of the available NeRF and 3D Gaussian Splatting models for room-scanned videos
- Streamed the AI-rendered images from the server to a client viewer
- Awarded 2nd place(최우수상) in the Co-Deep Learning Project, SKKU
Automated Summarization and Authoring Assistant for Product Review Data
Team Member | 2020.12 ~ 2021.02
- Industry-University Coop Project with Goodmorning Information Technology
- Finetuned KoGPT-2 on cosmetics review dataset crawled from various review sites, for review generation
- Finetuned KoBART on DACON news summarization dataset, for review summarization
PreVentra : AI-detected Virus Infection Prevention System
Team Member | 2020.04 ~ 2020.12
- Industry-University Coop Project with Goodmorning Information Technology
- Set up the system for training and inferencing using YOLOv4 model
- Processed real-time video streams and scheduled the conversion of recorded data into stats
- Awarded 3rd place(우수상) in the 2020 Sol·Jik Challenge, Ministry of Science and ICT
Lecture Projects and Others
- Speech Enhancement using AI and Improvement of its Baseline Performance, AI Project, SKKU, Spring 2024
- CLab - Interactive Clab and Lab Search Service using RAG, Introduction to Software Engineering, SKKU, Spring 2024
- Designing a Remote Couple Messanger <Dove>, Introduction to Human-Computer Interaction, SKKU, Spring 2024
- SpaceSwim Game Development Project, Computer Graphics, SKKU, Spring 2021
- PACMAN Game Development Project, Java Programming Lab, SKKU, Spring 2020
Personal
jazzy-dalpeng (working on progress)
- Jazz Guitar practice helper
- Integrates and syncs a canvas renderer and an audio player loop
- Built using SvelteKit (w. Svelte5) and custom HTML5 Canvas library
- Implements core system of game engine, including: Game loop, Component system, Linear algebra math library, Physically based rendering, Deferred rendering...
- Designed modern-web-frontend-like scripting features
- Built using WebGL2 and TypeScript
Research
Analyzing Character Relationship Shift Based on Action-Theory of Mind Trajectory
Jeong-won Kim, Teawoo You, Jong-hyun Lee, Yungyung Cheong
2025 한국소프트웨어종합학술대회 (KSC 2025)
Education & Work
Education
2025.09 ~ : Sungkyunkwan University, M.S. — INGLab
Major : Computer Science and Engineering
2019.03 ~ 2025.08 : Sungkyunkwan University, B.S.
Major : Computer Science and Engineering
Major GPA : 4.16 / 4.5
Participated Programs
- Toy Project Supporting Program
- Sungkyun SW Mentoring
- Co-Deep Learning
- SPARK 11st
- Industry-University Cooperation
Clubs
- Metaverse/XR Club MAV, joining as a committee member and a mentor, 2023 ~ 2024
- Algorithm/Problem Solving Club NPC, joined as a committee member, 2020 ~ 2021
- Classical Music Club Eumchiheon, 2019 ~ 2024
You can find my resources for personal study:
Work Experience
2023.07 ~ 2023.09: MUTI
- Interned as an Android app developer
- Developed an AR demo Android app, using Jetpack Compose and ArCore
2021.09 ~ 2023.04: ROK Air Force
- Served in Air Force Intelligence Information Systems Management Group as a software developer (대한민국 공군 정보체계관리단 소프트웨어개발병)
- Developed and managed multiple Air Force Intranet services, mainly using Spring, OracleDB, and Vue
Achievements and Certifications
Achievements
- 2024.01: 2nd place(최우수상) in Co-Deep Learning Project, SKKU
- 2020.12: 3rd place(우수상) in the 2020 Sol·Jik Challenge, Ministry of Science and ICT(과기부 주관 2020 솔·직 챌린지)
Scholarships
- 2023 Fall: Co-Deep Learning
Certifications
- 2024.08: 960 TOEIC score
Tech Skills
AI / LLM
- Designed and built LLM-based NPC Agent systems using OpenAI API
- Fine-tuned GPT-OSS-20B with Unsloth
- Trained, deployed, and integrated various AI models: YOLO, KoGPT-2, KoBART, BERT, NeRF, 3D Gaussian Splatting
Web/App Client
- Next.js (w. React), for many official web projects
- Vue, while working on ROK Air Force
- SvelteKit (w. Svelte), for many personal projects, including this page!
- Jetpack Compose for Android
- Unity for game development
Computer Graphics & XR
- OpenGL & WebGL2
- Android ArCore
Backend & DB
- Spring, Express(Node.js), FastAPI
- Oracle, MySQL, PostgreSQL
Kim Jeong-won
M.S. Student, Computer Science and Engineering, Sungkyunkwan University · INGLab
Interests: AI Agents ・ Society ・ World, Runtime AI for Games
Projects
Team
Shiba Games : Practical research and business planning on LLM-based NPC Agents
Team Leader | 2024.06 ~ 2025.02
- A project for SPARK 14th
- Explored runtime AI for game NPCs using LLM APIs — from research to startup planning
ShibaRPG — RPG system with interactive NPC conversations and actions
2024.11 ~ 2025.02
- Built an RPG where players can converse, battle, and trade items with LLM-driven NPCs
- Validated user immersion through open storytelling
ShibaMarket — Interactive commodity trading game with LLM-based agents
2024.06 ~ 2024.10
- Designed and developed a trading game that utilizes LLM Agent's expressiveness and judgment for NPC interactions
RealBang : 3D room generation AI models and viewer based on Novel View Synthesis
Team Member | 2023.05 ~ 2024.01
- A project for SPARK 11st and Co-Deep Learning
- Compared the performance metrics of the available NeRF and 3D Gaussian Splatting models for room-scanned videos
- Streamed the AI-rendered images from the server to a client viewer
- Awarded 2nd place(최우수상) in the Co-Deep Learning Project, SKKU
Automated Summarization and Authoring Assistant for Product Review Data
Team Member | 2020.12 ~ 2021.02
- Industry-University Coop Project with Goodmorning Information Technology
- Finetuned KoGPT-2 on cosmetics review dataset crawled from various review sites, for review generation
- Finetuned KoBART on DACON news summarization dataset, for review summarization
PreVentra : AI-detected Virus Infection Prevention System
Team Member | 2020.04 ~ 2020.12
- Industry-University Coop Project with Goodmorning Information Technology
- Set up the system for training and inferencing using YOLOv4 model
- Processed real-time video streams and scheduled the conversion of recorded data into stats
- Awarded 3rd place(우수상) in the 2020 Sol·Jik Challenge, Ministry of Science and ICT
Lecture Projects and Others
- Speech Enhancement using AI and Improvement of its Baseline Performance, AI Project, SKKU, Spring 2024
- CLab - Interactive Clab and Lab Search Service using RAG, Introduction to Software Engineering, SKKU, Spring 2024
- Designing a Remote Couple Messanger <Dove>, Introduction to Human-Computer Interaction, SKKU, Spring 2024
- SpaceSwim Game Development Project, Computer Graphics, SKKU, Spring 2021
- PACMAN Game Development Project, Java Programming Lab, SKKU, Spring 2020
Personal
jazzy-dalpeng (working on progress)
- Jazz Guitar practice helper
- Integrates and syncs a canvas renderer and an audio player loop
- Built using SvelteKit (w. Svelte5) and custom HTML5 Canvas library
- Implements core system of game engine, including: Game loop, Component system, Linear algebra math library, Physically based rendering, Deferred rendering...
- Designed modern-web-frontend-like scripting features
- Built using WebGL2 and TypeScript
Research
Analyzing Character Relationship Shift Based on Action-Theory of Mind Trajectory
Jeong-won Kim, Teawoo You, Jong-hyun Lee, Yungyung Cheong
2025 한국소프트웨어종합학술대회 (KSC 2025)
Education & Work
Education
2025.09 ~ : Sungkyunkwan University, M.S. — INGLab
Major : Computer Science and Engineering
2019.03 ~ 2025.08 : Sungkyunkwan University, B.S.
Major : Computer Science and Engineering
Major GPA : 4.16 / 4.5
Participated Programs
- Toy Project Supporting Program
- Sungkyun SW Mentoring
- Co-Deep Learning
- SPARK 11st
- Industry-University Cooperation
Clubs
- Metaverse/XR Club MAV, joining as a committee member and a mentor, 2023 ~ 2024
- Algorithm/Problem Solving Club NPC, joined as a committee member, 2020 ~ 2021
- Classical Music Club Eumchiheon, 2019 ~ 2024
You can find my resources for personal study:
Work Experience
2023.07 ~ 2023.09: MUTI
- Interned as an Android app developer
- Developed an AR demo Android app, using Jetpack Compose and ArCore
2021.09 ~ 2023.04: ROK Air Force
- Served in Air Force Intelligence Information Systems Management Group as a software developer (대한민국 공군 정보체계관리단 소프트웨어개발병)
- Developed and managed multiple Air Force Intranet services, mainly using Spring, OracleDB, and Vue
Achievements and Certifications
Achievements
- 2024.01: 2nd place(최우수상) in Co-Deep Learning Project, SKKU
- 2020.12: 3rd place(우수상) in the 2020 Sol·Jik Challenge, Ministry of Science and ICT(과기부 주관 2020 솔·직 챌린지)
Scholarships
- 2023 Fall: Co-Deep Learning
Certifications
- 2024.08: 960 TOEIC score
Tech Skills
AI / LLM
- Designed and built LLM-based NPC Agent systems using OpenAI API
- Fine-tuned GPT-OSS-20B with Unsloth
- Trained, deployed, and integrated various AI models: YOLO, KoGPT-2, KoBART, BERT, NeRF, 3D Gaussian Splatting
Web/App Client
- Next.js (w. React), for many official web projects
- Vue, while working on ROK Air Force
- SvelteKit (w. Svelte), for many personal projects, including this page!
- Jetpack Compose for Android
- Unity for game development
Computer Graphics & XR
- OpenGL & WebGL2
- Android ArCore
Backend & DB
- Spring, Express(Node.js), FastAPI
- Oracle, MySQL, PostgreSQL