Albert Chang
Applied AI Engineer
Hi, I'm Albert, an applied AI engineer at AIPLUX. I've worked on enterprise knowledge tools, educational software, and offline AI assistants for medical applications. Before I wrote software for classrooms, I worked in one.
Teaching and Classroom Software
I've taught children learning to program and adults learning cloud and AI concepts. At ViewSonic, I brought that teaching experience to reading-level adaptation and curriculum-aligned question generation. Both involved familiar teaching decisions: what students need to learn, how difficult the material should be, and how to assess their understanding.
AI Systems in Practice
These days, I work on language models and agents for organizations that need their knowledge systems to run on premises. That includes finding relevant information, managing conversation context, and checking whether a change actually produces better answers. I also work on the less visible parts: tracing, backend services, documentation, and connections to existing business software.
Experience
AIPLUX
AI Optimization Engineer · Jan 2026 – Present
As part of the team building a production AI platform for on-premises use, I focus on model evaluation and optimization, agent workflows, and conversation context.
- I built an automated evaluation pipeline covering more than 40 models to compare answer quality, response time, and deployment tradeoffs, and to detect regressions as the system changes.
- As the sole developer, I designed and built a GraphRAG recommendation and search system for 35,200 articles.
- I delivered an ERP integration that connects existing business workflows with AI agents through LINE and web interfaces, working directly with customer stakeholders on technical requirements.
ViewSonic
AI Engineer · Jul 2024 – Dec 2025
I helped take a classroom AI platform from proof of concept through MVP development. My work included algorithms that turn education theory into measurable student signals, real-time safeguards for off-task or toxic conversations, and evaluation workflows to check the system's responses.
I later refactored an internal AI sales assistant, reorganizing its agent workflows and improving documentation and tracing. The architecture cleanup and documentation reduced engineer onboarding time from two weeks to three days.
Taiwan Main Orthopaedic Biotechnology
AI Software Engineer · Dec 2023 – Jun 2024
I designed and built the AI foundation for an offline anatomy-table assistant, bringing together local language models, knowledge retrieval, and speech capabilities on dual RTX 4090 hardware.
The AI components reached the MVP stage. I also worked with Unreal Engine developers on the interfaces needed to connect AI outputs with 3D visualization.
JCTI
Software Engineer Intern · Sep 2020 – Jul 2021
I built a Python crawler and LINE Notify bot for public bicycle dispatch staff. It collected station availability and sent updates automatically, replacing their hourly manual checks.
Teaching & Research
My teaching, research, and early software work overlapped during my time in Kinmen.
Kinmen Employment Center, Workforce Development Agency
Sessional Lecturer · Mar – Jul 2023
I designed and taught an ICT and Google Cloud course for 30 adult learners. We covered cloud computing, machine learning, and language-model concepts through practical exercises for people with limited computing experience.
Kinmen Hepu Primary School
Substitute Teacher & IT Section Chief · Aug 2021 – Jul 2022
I taught information technology and science, designed programming and robotics activities, and maintained the school's IT infrastructure. During the pandemic, I led the school's transition to remote learning within half a day, helping teachers and students make the change.
Independent Teaching
Freelance Technology Education Lecturer · Sep 2019 – Mar 2023
I taught at primary schools across Kinmen, designing lessons in computational thinking, Scratch, App Inventor, and robotics. I adapted the activities and teaching pace for different ages and classrooms.
National Quemoy University
NLP Research · Aug 2020 – Jul 2023
I built a 50-million-character Chinese corpus about Kinmen's history and culture, using web scraping, OCR, speech-to-text, and data cleaning. I used it to analyze vocabulary and to fine-tune GPT-2 for controllable text generation. This research informed my master's thesis, and related work was published at IEEE ICCE-TW 2023.
Education
National Quemoy University
- Master's Degree in Computer Science and Information Engineering · 2021 – 2023
- Bachelor's Degree in Computer Science and Information Engineering · 2017 – 2021
Around Here
This site is where I share notes, side projects, and what I'm learning along the way. If you'd like to talk about AI systems, education, or a potential collaboration, feel free to get in touch.