Hi! I'm Jun Seong Pyo πββοΈ, a researcher based in Seoul, South Korea. My work centers on multimodal AI and LLM-based research β particularly in task-oriented dialogue, smart home AI agents, and spoken language understanding.
- Role: Integrated M.S.-Ph.D. in Graduate School of Data Science
- Affiliations: HOLI Lab @ Seoul National University
- Research Interests: LLM Agents | Task-Oriented Dialogue | Smart Home AI | Spoken Language | RAG
-
SpokenUS: A Spoken User Simulator for Task-Oriented Dialogue
Jonggeun Lee*, Junseong Pyo* (*Co-first), Jeongmin Park, Yohan Jo
ACL Rolling Review (ARR), March 2026 β Under Review -
SpeakerSleuth: Evaluating Large Audio-Language Models as Judges for Multi-turn Speaker Consistency
Jonggeun Lee, Junseong Pyo, Gyuhyeon Seo, Yohan Jo
ACL 2026 - Main -
SimuHome: A Temporal- and Environment-Aware Benchmark for Smart Home LLM Agents
Gyuhyeon Seo, Jungwoo Yang, Junseong Pyo, Nalim Kim, Jonggeun Lee, Yohan Jo
ICLR 2026 β Oral
- B.F. Sword: Blank Inference System with Optimized RAG and Knowledge Distillation
Junseong Pyo, Hyejin Bae, Kitae Kwon, Jina Kim, Woojung Song, Yeonseong Shin
HCI Korea 2025
Integrated M.S.-Ph.D. | Sep 2026 β Current | Seoul, South Korea
- Developing audio-visual memory in streaming ego-centric videos
Undergraduate Research Assistant | Jul 2025 β Aug 2026 | Seoul, South Korea
- Developed benchmarks for smart home LLM agent evaluation and training
- Researched spoken user simulators in task-oriented dialogue (TOD)
- Built a dataset pipeline for paper discovery agents with KISTI
Research Intern | Mar 2025 β Jun 2025 | Seoul, South Korea
- Developed a VLM-LLM smart home system for personalized user instructions
- Built LG-SHARE, a multi-turn evaluation benchmark for LG Smart Home Air Care
- Developed an internal web platform for GPU monitoring
Frontend Web Developer | Mar 2024 β Aug 2024 | Seoul, South Korea
- Implemented SMS server integration for user authentication and recovery
- QA testing manager for the National Election Commission's online voting system
Teaching Assistant | March 2025 - Aug 2026 | Seoul, South Korea
- Served as a Teaching Assistant for the AI+X: Deep Learning course
- Supporting undergraduate students with assignments and lab sessions.
| Year | Award | Organizer |
|---|---|---|
| 2025 | Student Achievement Award (86th Anniversary) | President, Hanyang University |
| 2025 | First Prize (2nd Place), Engineering Competition | Dean of Engineering, Hanyang University |
| 2024 | IC-PBL Contest (2nd Place) β LLM smart home system | LG Electronics |
| 2024 | App & Service Dev. (3rd Place) β "Korail Donghae" LLM app | Mayor, Busan Metropolitan City |
| 2023 | Encouragement Prize (4th Place) β ML-based timetable platform | Mayor, Daegu Metropolitan City |
- SimuHome β Smart home simulator integrated with LLMs for multi-turn, context-aware agent evaluation (β ICLR 2026 Oral)
- Project ALIVE β LG Electronics collaboration; natural language-driven smart home multi-chat system
- FORIF OFFICIAL β Club administration platform for Hanyang University programming club
Seoul National University, Seoul, South Korea Integrated M.S./Ph.D. Program, Department of Data Science, Graduate School of Data Science | Sep 2026 - Current Overall GPA: -
Hanyang University, Seoul, South Korea
B.I.S. Department of Information System | Mar 2023 β Aug 2026
Overall GPA: 4.19 / Major GPA: 4.18
- Programming: Python, C/C++, TypeScript
- Research: LLM Agents, Task-Oriented Dialogue, Benchmark Construction, RAG, NLP
- Languages: Korean (native), English (fluent)
I'm an researcher working at the intersection of NLP and human-centered AI. My focus is on how large language models can be rigorously evaluated and reliably deployed in real-world, context-dependent environments. I'm driven by the belief that better benchmarks and more realistic simulators are the foundation for trustworthy AI.
FORIF (Hanyang Programming Club)
Member (2023β2026), President (2024)
- Organized club events, increasing membership
- Managed algorithm competition (HSPC)