National Tsing Hua University
M.S. in Computer Science · GPA 3.75 / 4.3
Software, IT & AI Engineering · NTHU CS
M.S. in Computer Science · GPA 3.75 / 4.3
B.S. in Computer Science and Information Engineering · GPA 4.03 / 4.5
Coursework: Cloud Programming, Data Mining, Advanced Database Systems, Advanced Computer Architecture
IT Engineer Intern
Tool Software Engineer Intern
Co-authored a single-owner K/V residency design for exact Llama 3 attention on Alveo U50; member of the team awarded 1st Prize in the FPT 2026 Design Competition, Track B.
Built a five-stage RV32IMA core and configurable write-back cache in Chisel; verified 40 GEMM/cache configurations, 810 cache checks and 58 upstream ISA binaries using an independent ISA simulator and retirement lockstep.
Built a Python retrieval benchmark across 1,500 public-data questions, comparing keyword, dense, hybrid and graph-based methods with source attribution, latency profiling and failure analysis; implemented versioned retrieval contracts and 20 functional checks.
Reproduced a 15-stage lab LLM classification pipeline and built a six-condition context study over 241 unique utterances. Added paired error analysis and a Python/HTML review interface with sentence-level audio and HTTP Range seeking.
Automated a nested-KVM live-migration lab with QEMU, NFS and QMP; migrated a running Linux workload with 14 ms QEMU-reported VM downtime and built reproducible CPU, network and VM/microVM benchmarks.
Designed and implemented a hybrid spatial/graph database architecture using PostGIS and Neo4j, containerized with Docker. The project was selected by a Taichung clinic for a paid collaboration.
Built a document-grounded sustainability question-answering system with source traceability and prompt engineering; ranked 16th among 66 teams in Wattbot 2025.
Designed an XGBoost machine learning pipeline for bank alert account prediction; finished in the top 20% of 790 teams in AI CUP.
Implemented automated document question answering with LangChain, Llama 3.2 and ChromaDB, combining semantic retrieval with language-model generation.
Developed and verified hardware modules for a systolic-array deep learning accelerator on FPGA using Verilog, FPGA development tools and GTKWave.
Fine-tuned RoBERTa for six-class emotion classification with stratified validation and macro-F1 checkpoint selection; compared models, preprocessing, class weights and learning rates.
Built a text-mining workflow with TF-IDF, frequent-pattern mining and classifier comparisons; explored PCA, t-SNE and UMAP representations with interactive Plotly visualizations.
Extended a course scaffold with Pydantic transcript/video schemas and Gemini-based segment analysis, combining lecture transcription with visual insights.
Created reproducible PyTorch-to-ONNX export and analysis tools for ResNet-50, AlexNet and ViT-B/16; validated all three models and analyzed operators and parameter costs for hardware mapping.
TOEIC 920
GDSC NDHU core member; organized and hosted a technical meetup in May 2023.