CV
Curriculum Vitae of Yujin Bae.
Contact Information
| Name | Yujin Bae |
| Professional Title | M.S. Student, KAIST |
| yujinbae@kaist.ac.kr | |
| Location | Daejeon, South Korea |
| Website | https://jinyubae.github.io |
Professional Summary
I am interested in building multimodal large language models that genuinely understand the 3D world. My current work centers on human action understanding, leveraging 3D scene representations to predict and reason about human motion. Building on this, my broader goal is to strengthen robot intelligence with language and multimodality, enabling robots to act reliably and intelligently in the human-centered spaces we live in. My research interests span multi-modal LLMs, egocentric vision, human action anticipation, and human–robot interaction.
Education
-
2025 - Daejeon, South Korea
M.S.
Korea Advanced Institute of Science and Technology (KAIST)
The Robotics Program
- Advisor: Prof. Kuk-Jin Yoon (Visual Intelligence Lab)
-
2019 - 2025 Seoul, South Korea
B.S.
Ewha Womans University
Department of Software
- Summa Cum Laude (Highest Honors)
- Ranked 1st in class
-
2023 - 2023 Vermont, USA
Exchange Program
University of Vermont
Exchange Student
- Funded by Mirae Asset Exchange Scholarship
Honors and Awards
-
2025 Summa Cum Laude (Highest Honors)
Ewha Womans University
-
2022 Minister's Award, Ministry of Science and ICT — Blockchain Grand Challenge
Ministry of Science and ICT, Republic of Korea
-
2022 Dean's List
Ewha Womans University
Awarded 2019–2022.
Research Interests
Multi-modal LLMs: grounding language models in the 3D world, 3D-aware vision–language models
Egocentric Vision: reasoning about the world from a first-person view
Human Action Anticipation: predicting what a person will do next, human motion prediction
Human–Robot Interaction: letting robots assist people proactively
Research Experience
-
2025 - Daejeon, South Korea
Project Member
Visual Intelligence Lab, KAIST
“Core Technology Development for Autonomous Driving in Unstructured Off-Road Environments,” funded by Hanwha Aerospace. Real-world deployment target in unstructured off-road terrain.
- Core perception technology for autonomous driving in unstructured off-road environments
- Data augmentation and sensor fusion for robust perception
- Real-world deployment target in off-road terrain
-
2024 - 2024 Daejeon, South Korea
Research Intern (Full-time)
HCI Tech Lab, Graduate School of Culture Technology, KAIST
-
2022 - 2022 Daejeon, South Korea
Research Intern (Full-time)
Humanoid Generalization Lab, Graduate School of AI, KAIST
Languages
Korean : Native
English : Fluent (TOEIC 960)
Skills
Programming & Tools: Python, Git, Bash, LaTeX
ML / Scientific Computing: PyTorch, TensorFlow, PyBullet