Wesley Huang

AI RESEARCHER · DATA ENGINEER

Wesley Huang黃暐宸

Taipei, Taiwan · wesleyseawatch@gmail.com · +886 972-572-606

Statistics graduate and AI researcher with experience in agentic AI, multilingual NLP, machine learning, financial data, and business automation. I build reproducible research pipelines and production-oriented prototypes, with experience across NTPU, Acer, Eastspring Investments, and Winbond.

Download CV↓View research↘
Portrait of Wesley Huang
CURRENT FOCUS

Agentic AI · Multilingual ESG Compliance · LLM Evaluation · NLP

01

Education

National Taipei University

Bachelor of Science in Statistics · Graduated

  • GPA: 3.9 / 4.0.
  • Graduation average: 88.3 / 100.
  • Dean’s List Award, first place, Fall 2022.
  • TOEIC Gold Certificate: 865.
  • TOEFL Overall: 5 / 6.
02

Research

ROCLING 2025 · ORAL PRESENTATION · FIRST AUTHOR

RegComAgent & Multilingual Promise Verification

NTPU Intelligent Financial Innovation Technology Lab · Research Assistant

  • Developed RegComAgent, a language-neutral agentic AI framework for multilingual ESG regulatory-compliance verification.
  • Designed a reproducible multi-step reasoning pipeline and analyzed four systematic LLM failure modes.
  • Optimized an LLM sustainability-commitment extraction pipeline, improving processing efficiency by 50%.
  • First author of a ROCLING 2025 oral-presentation paper evaluating LLM performance on multilingual ESG promise verification.
Agentic AILLM EvaluationESGMultilingual NLP
Paper link — coming soonCode — coming soonDataset — coming soon

Financial ESG Data Engineering

Research collaboration with a University of Pennsylvania PhD candidate

  • Connected Python to WRDS APIs and processed 780,000 records from ISS ESG Analytics, CRSP, and Markit Securities Finance.
  • Conducted participation analysis, independence tests, normality tests, and statistical modeling with pandas, NumPy, SciPy, and statsmodels.
03

Industry Experience

AI & Cloud Engineer Intern

Acer Inc. · AFS

  • Developed an AI-driven sentiment-monitoring system with automated reporting, reducing daily analysis from 1.5 hours to 0.17 hours (89%).
  • Built an NLP competitive-intelligence MVP, saving 1.5 hours per person daily and reducing review time by 70%.
  • Implemented a wearable-device LLM chatbot MVP, reducing development rework by 30% through early validation.
  • Created a Power Automate procurement workflow, saving 40% of monthly administrative time.

IT Department Intern

Eastspring Investments · Prudential plc

  • Optimized a Python TDCC web scraper, reducing execution time by 50%.
  • Developed a Python customer-service email system with data anonymization, improving efficiency by 40%.
  • Implemented a chairman-approval workflow with Power Automate and Power Apps.

Data Optimization & AI Applications Intern

Winbond Electronics

  • Supported Azure OpenAI fine-tuning and developed a Flask AI assistant for SQL and naming tasks.
  • Built a BERT and Word2Vec course-recommendation system using OpenAI API, achieving 90% satisfaction.
  • Automated annual reporting with Python and Power BI, reducing processing from three days to two hours.

Data Analyst Intern

SeaWatch Co., Ltd.

  • Used GA3 behavioral data to optimize content and improve website performance by 15%.
  • Cleaned 10,000+ ocean-data records and performed time-series analysis in Python.
  • Deployed an API on Google Cloud Platform and built a Leaflet real-time tracking system.
04

Project Experience

Machine Learning Parts Prediction

Hotai Motor / Toyota · Project Member

  • Cleaned and analyzed 600,000 repair records and developed an automated ETL pipeline.
  • Developed an XGBoost model with mean absolute error controlled within 1.5 parts.
  • Maintained 95%+ precision through daily error monitoring, retraining, and A/B testing.
  • Detailed outputs are not public because the project is covered by a confidentiality agreement.

Awesome Taiwan Knowledge

g0v · Taiwan Zero School

  • Collected expert-validated educational data on Taiwan-specific knowledge and designed diverse question formats.
  • Built a dataset for model testing and evaluation; applied fine-tuning and RAG to Taiwan-focused language models.
Hugging Face Dataset ↗
05

Leadership

Leader

Google Developer Groups on Campus NTPU

  • Taught the 2025 Build with AI course on AI agents, authored materials, and led workshops and company visits.
  • Led more than 15 courses and workshops across programming, UI/UX, and AI applications.
  • Co-organized Taipei DevFest with 1,000+ attendees and managed content that reached 30,000 views.
  • Received NTPU’s Outstanding Academic Club recognition for the 2023 academic year.
06

Skills

AI / Machine Learning

LLM · Agentic AI · NLP · RAG · Fine-tuning · BERT · Word2Vec · XGBoost

Programming / Data

Python · pandas · NumPy · SciPy · statsmodels · SAS · R · SQL · ETL

Applications / Cloud

Azure OpenAI · OpenAI API · Flask · Google Cloud · Power Automate · Power Apps

Analytics / Visualization

Power BI · Tableau · GA4 · Seaborn · Leaflet · MySQL · Linux

07

Awards & Certifications

5Sensitive records · Verification available upon requestEMI TA, TOEIC, DevFest service, diploma, and employment record.