Seoul National University — HOLI Lab
Sep 2026 – Current
Ph.D. Student · Seoul, South Korea
- Developed Physical–API Interaction Environment for Robots
Junseong Pyo
Integrated MS-PhD Student in Graduate School of Data Science @ HOLI Lab, Seoul National University
Research Interests
I'm interested in multimodal AI and smart home agents grounded in speech and spoken language.
Speech and Spoken Language.
Most AI systems treat language as text, discarding the richness of how people actually speak. I'm interested in grounding agents in spoken interaction, where prosody, disfluency, and speaker identity carry meaning beyond the words.
Smart Home Agents.
The home is the space where people spend most of their lives — yet it remains one of the least explored frontiers in AI research. Home environments are dynamic and personal, accumulating time-series context that agents must reason over — not just respond to immediate commands. I work on benchmarks and evaluation frameworks that reveal where current AI agents fall short in this setting.
If you're interested in similar topics, feel free to reach out!
International Publications
[EMNLP]
Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, 2026
[ACL]
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics, 2026
[ICLR]
The Fourteenth International Conference on Learning Representations (Oral), 2026
Domestic Publications
[HCI Korea]
Proceedings of HCI Korea, 2025
Research Experience
Seoul National University — HOLI Lab
Jul 2025 – Aug 2026
Undergraduate Research Assistant · Seoul, South Korea
- Developed benchmarks for smart home agent evaluation and training
- Researched voice user simulators in task-oriented dialogue (TOD)
- Built a dataset for paper discovery agents
Education
Seoul National University
Enrolled: Sep 2026
Ph.D. Student, Department of Data Science, Graduate School of Data Science · Seoul, South Korea
Overall GPA: - / 4.3
Hanyang University
Mar 2023 – Aug 2026
B.I.S., Department of Information Systems · Seoul, South Korea
Overall GPA: 4.19 / 4.5 · Major GPA: 4.18 / 4.5
Work Experience
LG Electronics — AI Advanced Research Team
Mar 2025 – Jun 2025
Intern · Seoul, South Korea
- Developed a VLM-LLM smart home system for personalized user instructions and requests
- Built a web platform to monitor GPU usage and processes
- Built LG-SHARE, a multi-turn evaluation benchmark for LG Smart Home Air Care
Zkcrypto — SW Team
Mar 2024 – Aug 2024
Frontend Web Developer · Seoul, South Korea
- Implemented SMS server integration for user authentication and recovery
- QA testing manager for the National Election Commission's online voting system
Projects
LG ALIVE
2024.09 – 2024.12
Lead · Design · Frontend Development
- LG Electronics collaboration on a natural language-driven smart home multi-chat system
FORIF OFFICIAL
2024.01 – 2025.12
Lead · Design · Frontend Development
- Club administration platform for Hanyang University programming club
Awards
DIVE 2026
2026
2rd Place · Personalized public rental housing recommendation service · Mayor, Busan Metropolitan City
Hanyang University's Student Achievement Award
2025
86th Anniversary · President, Hanyang University
Hanyang University's Engineering Competition
2025
2nd Place · Dean of Engineering, Hanyang University
Information System's IC-PBL Contest
2024
2nd Place · LLM smart home system · LG Electronics
DIVE 2024
2024
3rd Place · "Korail Donghae" LLM app · Mayor, Busan Metropolitan City
Global Innovator Festa (GIF)
2023
4th Place · ML-based timetable platform · Mayor, Daegu Metropolitan City
Patents
English Problem Creation System Using Multi-Agents
2024
Patent · App. No: 10-2024-0109541 · Filed: August 16, 2024
Skills
Programming:
Python · C/C++ · TypeScript
Languages:
Korean (native) · English (fluent)