LEAD SOFTWARE ENGINEER 1 · CAMUS TECHNOLOGIESMS STUDENT · NATIONAL TSING HUA UNIVERSITY
01 / APPROACH
From a model’s potential to a system people use.
My work connects applied AI with the engineering that makes it useful: backend architecture, data isolation, deployment reliability, and thoughtful product integration.
I’m currently a Lead Software Engineer 1 (AI/ML) at Camus Technologies in Taipei and an MS student at National Tsing Hua University. Alongside product engineering, I research computational imaging at National Taiwan University.
02 / SELECTED WORK
Systems. Images. Useful outcomes.
A selection of production engineering, research, and independent projects.
PRODUCTION SYSTEM / TAIPEI
ALERT SIGNAL→REAL-TIME RELAY→PASSENGER DISPLAYS
ZERO-DOWNTIME REDEPLOYMENTS
01 / PRODUCTION ENGINEERING
Information that can’t wait.
Deployed and verified a real-time emergency-alert relay for Taipei Metro’s passenger information displays under a government contract.
Carried out zero-downtime production redeployments during overnight maintenance windows.
My contribution: deployment and verification of the relay, including production redeployments within overnight maintenance windows. The diagram is a conceptual illustration, not an internal system schematic.
COMPUTATIONAL IMAGING / NTU
PROJECTIONS → RECONSTRUCTION → CLARITY
02 / COMPUTER VISION RESEARCH
Seeing beyond the artifacts.
Researching SART-based 3D CT reconstruction and image-processing pipelines for volumetric analysis at NTU’s Digital Image Processing and Computer Vision Lab.
Developing U-Net models for shadow and illumination artifact removal in X-ray images to improve structural clarity for defect analysis.
PyTorchU-Net3D reconstruction
Research focus +
Two connected areas: reconstructing volumes from CT projections and improving X-ray image quality through deep learning. The visual is an illustration of the research domain, not an experimental result.
DEFECTQUEST / MULTIMODAL AI
LOCALISE / ASSESS / EXPLAIN
03 / APPLIED AI
Inspection with an explanation.
Built an SMT inspection system that analyses BGA X-ray images to detect per-ball solder defects and evaluate compliance.
DefectQuest brings together defect localisation, risk scoring, root-cause reasoning, and sustainability impact estimation for electronics manufacturing.
Multimodal AIX-ray analysisDefect detection
Project scope +
The system brings detection and interpretation into one inspection workflow. The solder-ball grid is illustrative; it does not represent a measured inspection result.
Production LLM integration, prompt caching, and tool-use design. Pluggable backend architecture across 28+ product modules, with database-level row-level security and multi-tenant isolation.
MAR — JUL 2026
Software Engineer Intern (AI)
Camus Technologies · Taipei
Contributed to a clinical imaging pipeline with a DICOM viewer and AI-assisted radiology reporting. Tested core workflows and resolved reliability issues.
JUN 2025 — PRESENT
Undergraduate Research Assistant
National Taiwan University · Computer Vision Lab
Computational imaging, SART-based CT reconstruction, and U-Net-based X-ray image enhancement.
SEP 2024 — DEC 2025
Undergraduate Research Assistant
National Tsing Hua University · Algorithms Lab
Designed and deployed StudyBuddy with cosine-similarity ranking, preference-based matching, and NTHU email access control.
04 / FOUNDATIONS
Technical range. Academic depth.
National Tsing Hua University
MS · Institute of Information Systems and Applications Sep 2026 — Present
BS · Electrical Engineering and Computer Science Sep 2022 — Jun 2026