Seongyong Kim
Connecting computer vision, AI, and robotics with the needs of construction sites.
Georgia Institute of Technology
Sensing Construction Dynamics (SeCD)
I am currently pursuing a Ph.D. in Civil Engineering at Georgia Tech, advised by Dr. Yong K. Cho. As a member of the Robotics and Intelligent Construction Automation Lab (RICAL), my research combines 3D perception, robotics, and AI agents to turn construction-site observations into actionable information grounded in site evidence.
At Seoul National University, I worked in the Spatial INformatics group under the supervision of Dr. Yongil Kim and Dr. Kiyun Yu. Following that, I gained experience as a GIS-focused startup entrepreneur, leading a government-funded project on business site selection.
Education
Ph.D. in Civil Engineering, 2021 ~
Georgia Tech, Atlanta, Georgia
MS in Computer Science, 2024
Georgia Tech, Atlanta, Georgia
MS in Civil Engineering, 2019
Seoul National University, South Korea
BS in Architectural Engineering, 2017
Seoul National University, South Korea
Interests
Construction Site Perception and Understanding
- 3D Reconstruction & Segmentation, Scan-to-BIM
- Agentic AI for Project-Level Information Integration
- Construction Robotics and Automation – Drones, Quadrupeds
- Human–Robot Interaction and Construction Safety
Projects
SiteQuery: AI Agents for Site Assessment
Connecting natural-language questions with floor plans and 3D site observations to provide evidence-grounded answers.
Field Data Collection & 3D Processing
Collecting data directly on fields and processing 3D pipeline to understand site conditions.
Human-Robot Interaction in Construction
Investigating how workers perceive and respond to robots in shared construction environments.
Virtual Construction Environments
Conducting interviews and surveys with construction practitioners using simulated scenarios to gather feedback on worker–robot interaction.
Vision-Based Work-Zone Safety
Detecting and tracking approaching vehicles to support intrusion alerts in highway work zones
UAV-Based Highway Asset Monitoring
Using UAV imagery and 3D reconstruction to assess roadside conditions and support maintenance planning.
Vision-based structural monitoring
Using camera-based measurements to monitor structural movement and deformation.
Mass Concrete Thermal Management
Evaluating temperature development and cooling strategies to support mass concrete construction planning.
![Indoor Reconstruction [Korea MOLIT, Jul. 2017 – Dec. 2020]](/project/2020-indoor/featured_hu53cef8c9ec1a8a63161258650392b915_2097013_550x0_resize_lanczos_2.png)
Indoor Reconstruction [Korea MOLIT, Jul. 2017 – Dec. 2020]
Floorplan Analysis: retriving the indoor structure from floorplans
![GIS- Consulting Location Allocation as an Adversory Researcher [Samsung Economic Research Institute, May–Jun. 2020]](/project/2020-seri/featured_hu93eb2c34708d457bd50dbb15bdcd8b06_122942_550x0_resize_lanczos_2.png)
GIS- Consulting Location Allocation as an Adversory Researcher [Samsung Economic Research Institute, May–Jun. 2020]
GIS consulting based on open source software (QGIS)
Startup- recommending starting locations to new business owners
Funded by Korea Ministry of SMEs and Startups; a MVP will be ready by Jan. 2021
![GIS- Updating Land Category in Cadastral Map Using Hyperspectral UAV [Korea LX, Jan.–Dec. 2019]](/project/2019-landcover/featured_hu3e488ba09ccb57c11b5229d114d3170b_871438_550x0_resize_lanczos_2.png)
GIS- Updating Land Category in Cadastral Map Using Hyperspectral UAV [Korea LX, Jan.–Dec. 2019]
Cadastral Map; Discrepancy Analysis; Land Update; Hyperspectral UAV; Land Classificaion; 3D-CNN; Integration Analysis
Startup- launching the web application on real estate in Korea
집현전, the web service for recommending real estate to customers and linking them to real estate agents with low broker fees.
Featured
Rethinking Point Sampling for 3D Semantic Segmentation: An Incidence-Aware Approach to Construction LiDAR
Keywords: 3D semantic segmentation; Terrestrial laser scanning; Point cloud sampling; Construction monitoring
Advanced Engineering Informatics, Under Review
SIP: Site in Pieces-A Dataset of Disaggregated Construction-Phase 3D Scans for Semantic Segmentation and Scene Understanding
Keywords: Construction Lidar Dataset; 3D Point Cloud; Semantic Segmentation; Digital Twins
Journal of Computing in Civil Engineering, In Production
Teaching & Professional Development
CETL8718 — Tech to Teaching Capstone
Fall 2026
Refined my teaching through an iterative cycle of classroom observation, feedback, instructional revision, and follow-up observation.
BC 2610 — Construction Technology I
Guest Lecturer · Fall 2026
Delivered a guest lecture on 3D reality capture and as-built reconstruction for undergraduate construction students.
CEE 6185 — Construction Automation
Teaching Assistant & Lecturer · Fall 2022, 2024–2026
Contributed to course module design, and assignment development, alongside lectures on construction automation.
OMSCS 7641 — Machine Learning
Teaching Assistant · Summer–Fall 2026
CEE 8813 — Construction Equipment and Methods
Lab Instructor · Spring 2024–2025
Led five weeks of hands-on lab sessions on construction simulation modeling.
CETL 8713 — Teaching and Learning in Higher Education
Spring 2024
Strengthened my foundation in learning principles, course design, inclusive teaching, and student assessment.
Gallery
