|
I am a Research Intern at RideFlux, where I work on autonomous driving.
I received my B.S. in Electrical Engineering from Kookmin University, where I worked with Prof.
Seongwon Lee on 3D/4D scene understanding.
Previously, I was a Research Intern at CARIAD, advised by Dr.
Xavier Timoneda.
Email / Google Scholar / Github / CV |
|
|
My long-term goal is to build spatially grounded world models for autonomous agents that can predict how the physical world may evolve under possible actions and make reliable decisions. I believe that understanding the world is most useful when it can inform an agent’s behavior. An autonomous agent does not need to model every aspect of its environment. It must capture the geometry, uncertainty, and dynamics that matter for the decisions it will eventually make. My work so far has explored this question through 3D/4D scene understanding, and I am increasingly interested in extending it toward predictive models that reason about the consequences of actions. Ultimately, I want to understand how perception and world modeling can be learned together with the decisions they are meant to support. |
|
A plug-and-play module that selectively corrects low-confidence appearance predictions.
Junho Kim, Jiseok Kim,
Seongwon Lee
International Conference on Control, Automation, and Systems (ICCAS), 2026 arXiv |
|
Multi-surface Ray occupancy modeling with Gaussian mixture model.
Junho Kim,
Seongwon Lee
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2026 arXiv |
|
Cross-view geometry grounding produces cleaner 3D occupancy with fewer Gaussians.
Junho Kim,
Seongwon Lee
IEEE International Conference on Robotics and Automation (ICRA), 2026 project page / arXiv / code |
|
|
|
Reflective vehicle reconstruction with disentangled diffuse and specular appearance.
Junho Kim,
Sangho Byeon,
Jiseok Kim,
Seongwon Lee
Team Lead of We-Meet Project (Fall 2025) |
|
Real-time autonomous driving perception and deployment on an embedded platform.
Minchan Jeong,
Junho Kim,
Dahyeon Ko,
Jaehyoung Park,
Donghwan Seo
Perception Engineer / ROS, Multi-View 3D object detection, V2X 3rd Prize, (top 3 out of 80 teams worldwide) |
|
|
 |
RideFlux, Seoul, South Korea
2026.08 - Present Perception Research Intern |
 |
Kookmin University, Seoul, South Korea
2019.03 - 2026.08 B.S. in Electrical Engineering Advisor: Prof. Seongwon Lee |
 |
CARIAD SE, Ingolstadt, Germany
2024.07 - 2024.12 Sensor Fusion Research Intern Advisor: Dr. Xavier Timoneda |
 |
SEA:ME, Wolfsburg, Germany
2023.07 - 2024.06 Out-Bound Program Advisor: Prof. Jongchan Kim |
 |
Republic of Korea Army, Gyeonggi-Do, South Korea
2020.12 - 2022.6 Mandatory Service |
|
Template stolen from Jon Barron and Yijia Weng . Thanks for stopping by :)
|