Tridge
AI-Powered News Classifier
LLM news classification and backend processing at Tridge.
My contributionDevelop the classifier, compare model versions, and improve processing algorithms.
Read case studyEJ Caliwag · Seoul, South Korea · UTC+9
AI Engineer. From models to working systems.
I work across applied AI, backend systems, and deployment. My experience spans LLM news classification at Tridge, computer vision at Tricubics, and edge–cloud research.
Open to remote, hybrid & on-site roles · Open to relocation
South Korea: F-series visa holder
Selected work
Systems I’ve worked on
Professional work and research, with the problem, my contribution, and the technical approach made explicit.
Try the portfolio assistant →A working demo with source-linked answers.
Tricubics
Unmanned Vending Machine System
Computer vision, tracking, and deployment for an unmanned vending service.
My contributionLead AI/software integration across detection, tracking, decision logic, and deployment.
Read case studyKumoh National Institute of Technology
Collaborative Edge-Cloud Face Recognition
Published research into face-recognition inference across edge and cloud systems.
My contributionCo-author the Sensors paper on collaborative edge–cloud inference.
Read case studyExperience
From research to delivery
- April 2025 – Present
Assistant Manager, ML Engineer
Tridge - Backend Department AI Team · South Korea
As an Assistant Manager on the AI Team within the Backend Department at Tridge, I am developing a news classifier leveraging large language models (LLMs). My work involves testing multiple versions of the model and implementing algorithms to enhance processing speed. I also contribute to improving the overall performance and efficiency of company systems.
- March 2024 – March 2025
Lead AI/Software Engineer
Tricubics - Research & Development (R&D) · South Korea
Oversee end-to-end service management for the unmanned vending machine system, including team progress tracking and scheduling. Designed and maintained DevOps processes using Docker and Kubernetes to support system updates and modifications. Applied YOLOv8 for object detection, optimizing performance with TensorRT; used MobileNetV3 for classification tasks. Implemented object tracking algorithms based on ByteTrack and DeepSORT. Developed a decision-making algorithm leveraging a 'frame zone' approach to enhance system intelligence. Integrated ZeroMQ for lightweight messaging protocol support. Added optional MongoDB integration for storing external transaction data.
- October 2022 – February 2024
AI Engineer
Data Design Engineering · South Korea
Image segmentation using U-Net architecture. Collaborate on writing algorithms as well as training a YOLOv5 and YOLOv7 model for book spine detection. Develop, train, and deploy a face detection model to an embedded board developed by KT&C. System automation and data pipeline handling using Apache Airflow, Docker, and AWS. Writing of Python scripts for crawling purposes. Database management including PostgreSQL and MongoDB.
Research
Published work
Let’s talk
Building an AI product or platform?
I’m interested in Applied AI, Backend, and Full-Stack AI engineering roles. Based in Seoul and open to relocation.