Visionary Engineering Leader with 8+ years of executive experience directing a high-performing 15-member R&D team in AI/ML, 3D vision, and SLAM fields
Principal AI/Vision Engineer & Tech Lead with 20+ years of extensive research expertise and deep technical knowledge spanning computational learning theory (machine/deep/reinforcement learning), artificial intelligence, computer vision, robotics, language processing, and pattern recognition
Proven Track Record of Productization: Successfully transformed a stagnant 4-year R&D project into a market-ready, world-class intraoral scanner (launching July 2026), elevating scanning accuracy to sub-100 um precision
Strategic Technical Contributor: Secured ₩5B in Series A funding and successfully won the Scale-up TIPS national project (₩1.2B budget) as a key technical architect
Cultivator of "Team-First Culture": Spearheaded communication-led collaboration frameworks and agile transformation (2-week team demos & Kanban boards), maximizing team synergy and technical velocity
Strong, Multidisciplinary Background in mathematics and mechanical systems, including machine science and control theory, enabling seamless hardware-software integration for complex intelligent systems
Ph.D., Robotics, Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Republic of Korea
Advisor: Hyun Seung Yang (Computer Science)
Research Areas: Statistical/Machine Learning, Reinforcement Learning, Manifold Learning, Computer Vision, Pattern Recognition, and Artificial Intelligence
Dissertation: "Human Activity Recognition Based on Bayesian Learning and Directional Statistics" (in Korean)
M.Sc., Mechanical Engineering, Yonsei University, Seoul, Republic of Korea
Advisor: Yong Je Choi (Mechanical Engineering)
Research Areas: Kinematics (Kinematic Geometry, Screw Theory), Dynamics, Control Theory, and Mechanism Design
Dissertation: "Vibrational Analysis of a Multi-Body System via Screw Theory" (in Korean)
B.Sc., Mechanical Engineering, Yonsei University, Seoul, Republic of Korea
Advisor: Kang-Rae Cho (Mechanical Engineering)
Lab: Applied Fluid Mechanics Laboratory
Dissertation: "Fluid Field Analysis around an Isolated Airfoil using PHOENICS" (in Korean)
Director, Inno3D Co., Ltd., Seoul, Republic of Korea
Engineering Leadership & "Team-First" Culture Transformation
Directed a cross-functional team of 15+ 3D vision, AI, and software engineers, shifting the organizational paradigm toward a goal-oriented collaboration framework based on single-line goals, desired outputs, and hierarchical structures
Implemented Scrum-like development methodologies, introducing 2-week team demos and Kanban boards to streamline cross-team communications and accelerate production velocity
Served as a core technical contributor to secure ₩5B in Series A funding across two rounds and won the prestigious Scale-up TIPS national R&D project (₩1.2B budget)
Supported the technical documentation, software validation, and QA compliance required to secure domestic KFDA and European CE medical certifications, with US FDA and Japan PMDA approvals ongoing
World-Class 3D Vision & SLAM Algorithm Productization
Spearheaded the commercialization of an intraoral 3D scanner, launching the company’s first commercial product in July 2026 after 4+ years of initial development stagnation
Elevated scan accuracy to less than 100 um, a world-class benchmark essential for the dental prosthesis fabrication process, securing the first commercial contracts at a recent international exhibition
Architected highly robust, environment-agnostic 3D point cloud registration algorithms utilizing C++, PCL, Open3D, and OpenCV, achieving 90%+ scanning performance compared to global market leaders
High-Performance AI/ML Medical Image Pipeline Design
Oversaw the R&D of Python & PyTorch-based instance segmentation and object detection approaches optimized for intraoral 3D data, achieving an F-score of 0.8 or higher
Optimized real-time data processing pipelines using CUDA and TensorRT, drastically reducing runtime latency from hundreds of ms to single-digit ms to meet chairside diagnostic requirements for live surgical and dental environments
Research Director, SiliconMinds Inc., Seoul, Republic of Korea
Document Processing Systems (DocuSense)
Developed document processing systems based on large-scale multimodal transformer models to process a variety of documents including books, forms, receipts, and tables
Developed intelligent document understanding systems using pretrained large language models and their fine-tuned models to extract useful information from documents
Text Processing Systems (TextSense)
Developed text processing systems based on RNN or transformer models to detect and recognize text in a variety of document and scene images
Developed Jamo-based Hangeul(Korean) recognition models based on attention mechanism
Motor Fault Diagnosis Systems (MotorSense)
Developed fault detection and diagnosis systems based on deep learning (CNN and RNN) and advanced signal processing technology
Automatic License Plate Recognition System
Developed a highly accurate license plate recognition system based on object detection and synthetic generation techniques
Team Leader, Anditek, Co., Ltd., Goyang, Republic of Korea
Development of Control Algorithms to Improve Accuracy and Stability of Linear Motor Systems
Developed control algorithms for high-precision positioning based on space vector modulation
Team Leader, AnyCasting Co., Ltd., Seoul, Republic of Korea
Development of a Defect Inspection System for Non-Destructive Testing (NDT) based on X-ray Images
Designed image processing algorithms to detect and localize defects of aircrafts
Developed a framework to deal with large-size, high-resolution X-ray images
Development of a Volume Data Reconstruction and Visualization Software Framework for Medical Applications
Developed a 3D volume rendering system based on VTK library
Project Leader, Cubictek Co., Ltd., Seoul, Republic of Korea
Development of Robot Off-Line Programming (OLP) Systems, V-Robot
Designed and implemented a flexible analysis framework for kinematics and dynamics
Developed algorithms for motion planning and trajectory blending from scratch
Supported various robot programming languages based on a general-purpose intermediate language
Integrated realistic 3D simulation and collision detection
Successful Application to Industrial Robots
Developed OLP systems for Cartesian, SCARA, and articulated robots of Hyundai Heavy Industries (HR series), Samsung Electronics (FARA series), Robostar, and Daewoo Heavy Industries
Development of an Integrated Simulation Software System, V-World
Designed and developed a general-purpose simulation engine based on discrete event system specification (DEVS) formalism
Integrated all simulation products (V-Robot, V-HPS, V-PLC, V-ELEQ, V-Meca) into a single framework for virtual manufacturing
Development of a Virtual Surgery System
Developed a software framework to reconstruct human organs from CT images and to visualize them
Designed a surface reconstruction algorithm based on marching cubes algorithm
Staff, Daewoo Heavy Industries Co., Ltd., Changwon, Republic of Korea
Development of a Machining Line for Automotive Engine Blocks of Daewoo Motors
S/W Consulting
Korea Hydro and Nuclear Power (KHNP) Central Research Institute
Conducted optimization of safety analysis software system
Korea Electric Power Corporation (KEPCO) Research Institute
Conducted software restructuring and refactoring of safety analysis software system for nuclear power plant
S/W Engineering
S/W Analysis and Design
S/W Re-structuring and Refactoring
S/W Acceleration and Optimization
Assistant Professor, Intelligent Robot Engineering, Mokwon University, Daejeon, Republic of Korea
Lectured in Introduction to Robotics, Design and Application of Microrobot, and Engineering Mathematics Basic
Lecturer, Intelligent Robot Engineering, Mokwon University, Daejeon, Republic of Korea
Lectured in Introduction to Robotics, Robot Structure Analysis, Walking Robot Control, Design and Application of Microcontroller, and Engineering Mechanics
Research Scientist, Center for Robotics Research, Korea Institute of Science and Technology (KIST), Seoul, Republic of Korea
Development of Image-based Phenotyping Technologies for Key Plants
Developed leaf/stem identification algorithms based on graph-theoretical approach and integer programming
Developed advanced digital phenotyping techniques based on multi-view images acquired from visible, near infrared (NIR), hyperspectral, and thermal image sensors
Developed plant structure extraction algorithms based on CNN and RNN
Post-doc, Intelligent Robotics Group, NASA Ames Research Center, Moffet Field, California USA
Safe Autonomous Systems Operations (SASO) Localization
Proposed a new integrated sensor model for image sensors like LIDAR, RADAR, sonar, and camera in order to facilitate sensor fusion
Designed a Bayesian-filtering-based framework for robust localization and mapping for UAVs based on multi-sensor data fusion in GPS-denied environments
Developed a framework for multi-sensor data fusion systems using various sensors such as LIDAR, RADAR, sonar, camera, and IMU
Built up a big picture for multi-sensor data fusion through extensive literature survey
Robust Asteroid Detection Algorithm (RADA)
Designed algorithms based on statistical inference using F-statistic and ANOVA and robust estimation to simultaneously detect moving objects and their trajectories
Prototyped an intelligent software to detect near-earth asteroids (NEAs) based on temporal snapshots of pixel data of sky
Developed a simple framework for performance evaluations of other algorithms
Developed test cases for software verification and validation and successfully applied an automatic build system to RADA using a couple of languages and several external libraries and tools
Ph.D., Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Republic of Korea
Real-time and Robust Vision-based Gesture Recognition System for Smart Media
Designed and implemented gesture recognition systems based on 2D image sensors for smart media such as smart TVs and smart phones
Developed and applied robust gesture recognition algorithms that are less vulnerable to variations in human, location, pose, and illumination
Designed a set of gestures to take recognition performance and ease of operation into consideration
Designed a flexible software framework that can be extended to various software and hardware platforms
Oil Spill Protection (OSP) Robots
Devised a two-loop control scheme for local positioning by visual servoing and global position by GPS
Designed a robust tracking and docking algorithm based on reinforcement learning and disturbance rejection in highly uncertain environments
Implemented autonomous surface vehicles (ASVs) including hardware and software systems
Designed embedded and Windows applications for sensor fusion of IMU, GPS, and camera via wireless network and I2C
Robot Grand Challenge 2009 in Korea
Implemented a mobile humanoid robot called AMIGO for the challenge, and then extended to a building tour-guide robot in combination with augmented reality (AR)
Designed a software framework for localization, navigation, speech recognition, object detection, and object recognition, etc.
Designed an error-tolerant software architecture to independently fulfill various S/W and H/W functions
Designed and integrated hardware including mechanical and electric parts: mobile platforms, a 6-DOF arm, various sensors, and controllers including various communication channels such as I2C, SPI, and socket
Designed embedded and Windows applications for sensor fusion of IMU, sonar, IR camera, and color camera
Extensively studied practical technologies for recognition of face, object, button, and elevator, etc.
Software Development for Healthcare Robots
Designed and implemented a common robot software middleware platform for target healthcare robots (chair-bot, ride-bot, and lifecare-bot)
Developed an inference engine for body and emotional states based on several bio-signals using ensemble learning and SVM
Built-up a communication framework for interface with a variety of bio-signal sensors like ECG, temperature sensor, and blood pressure sensor
M.Sc., Yonsei University, Seoul, Republic of Korea
Development of Machining Tool based on Stewart Platform
Designed and implemented a machining tool based on Stewart Platform in order to precisely manufacture components and parts with free surface and exert enough power
Also designed and implemented a control system based on nonlinear control algorithm for complex in-parallel robot systems
Development of 6-DOF Haptic Joystick using In-Parallel Mechanism
Developed a haptic master device for 6-DOF Schilling robot operating in nuclear atomic reactors
Designed the general-purpose master device as in-parallel mechanism to control all the serial and in-parallel slave robots
Adopted a master-slave system and implemented a 3D graphic simulation on workstation for tele-operation
Plant Area Extraction System and Method Based on Deep Learning and Connectivity Graphs, Junsik Kim and Sang-Wook Lee
Korea Patent No. 102131243
System and Method for Plant Leaf Identification, Sang-Wook Lee and Junsik Kim
US Patent No. US10679096B2, Korea Patent No. 101956158
Engine Assembly Inspection System using Acceleration Signals and a Device Measuring the Signals, Tae-Hwi Lee and Sang-Wook Lee
Korea Patent No. 100997897
Certified Information System Auditor (CISA), Information Systems Audit and Control Association (ISACA)
Project Management Professional (PMP), Project Management Institute (PMI)
KAIST Scholarship, Korea Advanced Institute of Science and Technology (KAIST)
Entrance Scholarship, Yonsei University