AI systems
LLM workflows, multimodal evaluation, event-driven applications, and research prototypes.
AI SYSTEMS · MACHINE LEARNING · COMPUTER ARCHITECTURE
I build AI systems that connect models, software, and real-world signals.
I am a Computer Science master's student at National Dong Hwa University in Taiwan, building research prototypes and practical systems across multimodal AI, event-driven applications, computer vision, and computer architecture.
- Open to internships and collaboration
- Taiwan
- M.S. Computer Science
ABOUT
From research questions to working systems.
My projects tend to start at the boundary between fields: an AI commentator driven by live game events, a motion-feedback pipeline built from pose estimation, or a network-on-chip study shaped around emerging memory workloads. I enjoy turning those ideas into testable architectures, code, experiments, and documentation.
Systems research
Computer architecture, network-on-chip simulation, performance analysis, and hardware-aware design.
Engineering practice
Python, Git, Linux, technical writing, testing, and open-source collaboration.
SELECTED WORK
Systems built to be tested, explained, and improved.
VIDEO AUTOMATIONV1 COMPLETE
Project Forge — Multi-Format Sports Video Pipeline
A reproducible Python and FFmpeg workflow that turns one selected training clip into branded 16:9, 9:16, and 1:1 MP4/GIF deliverables.
View case study →AI SYSTEMSACTIVE
Neko.exe — Event-Driven AI VTuber
A real-time character system that turns live game events into context-aware commentary using an LLM, conversation memory, and an event routing pipeline.
Case study in progressBACKENDPUBLIC REPO
Gaming Intelligence SaaS PoC
A FastAPI proof of concept for live-ops data ingestion, subscription-tier enforcement, secure media delivery, and asynchronous API access.
View repository ↗AUTOMATIONPUBLIC REPO
AI Print & Framing Automation
A programmatic high-resolution layout engine for scalable art-print production, dynamic mat proportions, and serverless image workflows.
View repository ↗COMPUTER VISIONRESEARCH PROTOTYPE
Figure Skating AI Coach
A phase-aware motion-analysis pipeline using pose estimation and dynamic time warping to turn skating movement into structured feedback.
Documentation in progressRESEARCH
Working across signals, architecture, and intelligent workloads.
01
MASTER'S RESEARCH DIRECTION
PIM-Aware Coherent NoC
Studying congestion onset and adaptive control in coherent on-chip networks supporting processing-in-memory workloads, with a small synthesizable digital PIM primitive.
02
SCIENTIFIC MACHINE LEARNING
Plant Electrophysiology Modeling
Exploring latent nonlinear dynamics and external stimulus coupling in plant electrical signals using sparse system identification and topology-aware analysis.
03
AI REASONING
NeuroGolf / ARC Program Synthesis
Building and evaluating programmatic transformations for abstract visual reasoning tasks, with an emphasis on reproducible candidate generation and validation.
OPEN SOURCE
Small changes, shipped through real review.
CARDANO FOUNDATION2 MERGED PRs
Contributor to cardano-org
Contributed production fixes and technical documentation improvements through the full GitHub review workflow.
LET'S CONNECT
Interested in AI systems, research engineering, and open-source collaboration.
The best way to reach me is through LinkedIn or GitHub.