KAIST · Kim Jaechul Graduate School of AI

Junhyeong Jeon

also known as Walter Jeon

I am an incoming M.S. student at the Kim Jaechul Graduate School of AI (KAIST AI), starting in August 2026 under Prof. Dongsu Han at the INA Lab. I am completing my B.S. in Computer Science at KAIST (Korea Advanced Institute of Science and Technology), with a double major in Electrical Engineering and a minor in Science & Technology Policy. I am broadly interested in systems research — particularly LLM inference optimization, distributed infrastructure, and multi-agent systems.

Education

Korea Advanced Institute of Science and Technology (KAIST)
Aug 2026 – Present (incoming)
Degree
M.S. in Artificial Intelligence
Admitted · entering Aug 2026
Advisor
Prof. Dongsu Han · INA Lab
Korea Advanced Institute of Science and Technology (KAIST)
School of Computing · Daejeon, South Korea
Mar 2020 – Aug 2026
Degree
B.S. in Computer Science
GPA / Credits
3.81 / 4.3 · 200 credits
Double Major
Electrical & Electronic Engineering
Minor
Science & Technology Policy (STP)

Research Interests

LLM Inference Systems
Efficient serving, speculative decoding, and hardware-aware optimization for large language models.
🌐
Distributed Systems & Infrastructure
Scalable, fault-tolerant infrastructure for cloud-native and datacenter-scale workloads.
🤖
Multi-Agent Systems
Coordination, communication, and system-level efficiency for collaborative AI agents.

Research Experience

LLM Inference Optimization & Agentic AI Systems
INA Lab, KAIST · Prof. Dongsu Han
Undergraduate Researcher → Incoming M.S. Student (Aug 2026)
Working on system-level optimization for agentic AI pipelines. Focusing on speculative decoding — investigating how to select and adapt decoding strategies based on workload characteristics to maximize inference efficiency across diverse agentic tasks.
mmWave-Based Weapon Detection System (Compressive Sensing)
CyPhy Lab, KAIST · Prof. Jun Han
Developed a weapon detection system using mmWave radar with metasurfaces. Optimized Frequency Selective Surface (FSS) unit-cells to enhance sensing efficiency and imaging reconstruction performance.
Exploration of Technology-Driven Security Challenges & Global Security Collaboration
FINST&P, KAIST · Commissioned by MOFA
Researcher
Conducted focused analysis on AI and cybersecurity as emerging strategic domains. Examined technological trajectories and policy approaches of the US, China, and the EU, and synthesized comparative insights to evaluate security implications and develop actionable strategic recommendations.
Exploration of Emerging Technologies & International Security Issues
FINST&P, KAIST · Commissioned by MOFA
Researcher
Conducted case-based research on emerging technologies and their implications for international security. Investigated policy responses across AI, cybersecurity, quantum science, space technology, battery systems, and critical minerals. Synthesized global technology trends and proposed strategic considerations for national security planning.
Effects of Medical School Quota Policy on STEM Human Resources
KAIST URP Program
Project Lead
Led empirical research examining how national medical school quota policies influence the long-term distribution of STEM human resources in South Korea. Designed the research framework, constructed datasets, and conducted quantitative analysis to assess policy implications on higher-education and workforce trends.

Publications & Presentations

Performance Improvement of Scene Text Detection Using Randomized Preprocessing and Efficient Backbone Utilization
KCC 2022 (Korea Computer Congress)
Effects of Medical School Quota Policy on STEM Human Resources
National STS-STP Undergraduate Conference 2025
Examining the Sewol Ferry Disaster from the Bottom Up
Disaster School, KAIST Graduate School of Science and Technology Policy

Projects

Zero Factory — Zero-Waste Lifestyle Mapping Application
Tech4Impact, KAIST
Project Manager · AI Engineer · DevOps Engineer · Team of 5
Led planning and execution of a mobile platform that maps zero-waste stores and verifies reusable-cup usage via AI vision. Developed an image classification model (ResNet + YOLO) for reusable vs. disposable containers, and built a production-grade deployment pipeline using Docker, GitHub Actions, and AWS EKS.
Python PyTorch FastAPI Kotlin Docker Kubernetes (EKS) GitHub Actions

Teaching Experience

CS101: Introduction to Programming
School of Computing, KAIST
Teaching Assistant
Teaching Assistant for KAIST's core introductory programming course across multiple semesters.
Special Lectures for High School Qualification Exam
Daejeon Seo-gu Out-of-School Youth Support Center
Teaching Mentor (Volunteer)
Ongoing volunteer teaching mentor for youth outside the formal school system. Recipient of the Best Mentor Award (2023, 2024).
Information Security for AI
SW Academy, KAIST
Served as the main instructor, designing and delivering curriculum on AI-oriented information security topics.
FLOW Camp — Coding Mentor
Center for Gifted Education, KAIST
Software Project: Face Tagging
Center for Gifted Education, KAIST
KAIST Science Outreach Program
Center for Gifted Education, KAIST
Informatics Subject Educational Volunteer
Jeolla High School

Lectures

EBS SW 2026 →
Common sessions (OT) and mentoring schedules with video meeting links for 7 teams.
Generative AI Special Lecture →
Daedeok Software Meister High School
Special lecture series on generative AI. Course materials hosted on Notion.
Information Security Special Lecture →
Daedeok Software Meister High School
Special lecture series on information security. Course materials hosted on Notion.