-
AI Engineer · E-Ranyx
Responsible for technical direction of AI-driven software, including large language models and computer vision in client platforms. Delivered an enterprise integration that reduced processing time by 80% and cost by 50%.
-
AI Engineer · DHT Solution
Designed and deployed machine learning systems. Evaluated LLM applications on a dataset of more than 10,000 examples and improved response accuracy by 90%.
-
AI Intern · Claireye Intelligence
Contributed to computer vision and LLM product development, including image recognition, collaboration with research and development on web and application interfaces, and frontend stability informed by user feedback.
-
AI Edge Course · ITRI
Designed and delivered edge-AI instruction for industry engineers, covering Linux, DeepStream SDK, Docker, and Azure IoT Edge on NVIDIA Jetson.
AI Engineer
ML · LLM · Computer Vision
Yi-Ho Chang
I develop and deploy machine learning systems for operational environments, including real-time computer vision, large language models, and containerized applications on edge devices.
MEng, Computer Science (AI minor), Oregon State University. AI Engineer at E-Ranyx, responsible for production AI architecture and for taking models from research through deployment.
About
Professional profile
I am an AI engineer and a Master of Engineering student in Computer Science at Oregon State University. My work focuses on machine learning, computer vision, large language models, and systems that must operate outside laboratory conditions, including remote and edge deployments.
As AI Engineer at E-Ranyx, I am responsible for how AI is implemented in the company’s platforms and for guiding work from research through production. My primary expertise is at the intersection of models and product systems: real-time vision (CNN, YOLO), generative and LLM pipelines including RAG, and the surrounding containerized software.
I emphasize architectures that can be documented, interfaces that expose rather than conceal complexity, and results that can be measured—including 93% weed-detection accuracy, a 90% improvement in LLM response accuracy, and an 80% reduction in client processing time. I seek roles in which I own features from initial investigation through production release.
Experience
Professional experience
Professional appointments in applied AI, including industry positions, internships, and instructional work. A formatted résumé is available as a PDF.
Download Resume (PDF)Projects
Selected projects
The following case studies cover multimodal interaction, computer vision, and production software platforms. Each entry documents context, method, and results.
Jun 2026 – Present
Interactive AI Avatar
An end-to-end multimodal avatar system. Speech, text, and visual input are processed through a low-latency interface and API gateway, then ASR, NLP, and a multimodal LLM, and returned as synchronized speech and animation.
- ASR
- Multimodal LLM
- TTS
May 2026 – Present
LINE Official Account Manager
A self-hosted LINE Official Account management system that unifies inbound messaging, human handover, automated replies, knowledge retrieval, and draft generation. Messages are transmitted only after explicit confirmation.
- LINE Messaging API
- Multi-LLM routing
- RAG
Jan 2026 – Mar 2026
Precision Agriculture Weed Detection
An object-detection pipeline for identifying weeds in high-resolution drone imagery of perennial ryegrass and tall fescue, using YOLO26 and a controlled study of tiling and input resolution.
- YOLO26
- Drone imagery
- Python
Education
Education
Academic training relevant to the design and deployment of AI systems.
-
Oregon State University
Sep 2025 – Dec 2026 · Advanced AI, Computer Vision, Machine Learning
-
National Taichung University of Education
Sep 2021 – Jun 2025
Certificates
Course credentials
Completed coursework in generative artificial intelligence and machine learning.
Contact
Contact
Please write with inquiries regarding positions, collaboration, or technical discussion. Correspondence is reviewed directly.