About Me

AI / LLM Engineer

Taipei, Taiwan

• Led development of 3+ production AI agents (Agentic RAG, multi-turn conversational AI, document review workflow), driving the acquisition of the company's first paying customers
• Established end-to-end LLM evaluation pipeline: offline testing with LLM-generated QA datasets and LLM-as-a-Judge, improving RAG answer accuracy from 0.65 to 0.80; online production monitoring via LangSmith reducing error rate from 5% to <1%
• Applied Deep Agent architecture and Context Engineering across production agents with dynamic tool-call limits; combined with prompt caching to reduce API costs by 60-75%
• Built purpose-built AWS infrastructure for AI agents with CDK: automated multi-jurisdiction data pipelines and agent file system, enabling agents to autonomously operate on uploaded documents

Machine Learning Research Intern

Taipei, Taiwan

• Researched Self-Supervised Learning for feature extraction on structured financial time series data (10M+ transaction records)
• Designed SSL approaches combining denoising autoencoder with contrastive learning, achieving 91% improvement in feature utilization and 10% boost in downstream model performance
• Delivered two research presentations on findings

M.S. in Information Management and Finance - Data Science Track

GPA: 4.27 / 4.3

• Academic Excellence Award Recipient
• Coursework: Machine Learning, Deep Learning, Data Mining, NLP

B.S. in Transportation and Communication Management Science

GPA: 3.99/ 4.3

IEEE Transactions on Knowledge and Data Engineering (TKDE) | SCI Q1

• Achieved 10% improvement in recommendation accuracy using Transformer-based cross-domain modeling
• Reduced demographic bias by 70% through novel transfer and contrastive learning techniques