EXPERIENCES
Senior Product Owner / Data Analyst
Supernova Technology | Chicago, US
- Took end-to-end ownership of the CoreBanking platform in Chicago, establishing structured workflows and operating standards that transformed a high-escalation, production-critical system into a stable product and reduced investigation and resolution time by ~50%.
- Led the productization and automation of decade-long manual financial workflows (APR updates and FIS rate maintenance) by defining system requirements and API integrations, eliminating 100% manual intervention and significantly reducing recurring operational risk.
- Designed standardized data workflows and the CoreBanking Integration Playbook, centralizing domain knowledge, reducing onboarding and integration ramp-up time by ~70%, and removing reliance on tribal knowledge.
- Owned the Money Movement configuration system end-to-end as a Product Owner, defining product requirements, designing workflows and system behavior, and delivering an integration-ready platform aligned across engineering, UX, and business teams.
- Applied SQL- and Python-driven analysis to booking logic, transaction flows, and data synchronization issues, improving defect detection efficiency by ~25% and enabling cleaner engineering handoffs and more reliable production systems.
- Acted as the single-threaded owner and decision-maker across Chicago and offshore teams, driving roadmap execution, cross-functional alignment, and delivery discipline, and becoming the trusted authority for CoreBanking systems among senior leadership and enterprise stakeholders.
Data Analyst Intern
Taelor | San Francisco, US
- Built automated analytics pipelines to maintain core business dashboards (Churn, Refund, Fulfillment, Styling) by integrating Python with Google Sheets API, Shopify, and BigQuery, ensuring reliable and real-time visibility into product performance.
- Defined and analyzed user engagement cohorts and funnels for preview messages and utilization behavior using SQL and Python, evaluating 3,000+ records to identify key engagement drivers and support data-driven product decisions.
- Designed executive-ready dashboards and visual reports that translated behavioral data into actionable insights, including persona segmentation and SMS response performance.
- Analyzed promotion campaigns and optimized targeting strategies, achieving a 72.22% increase in new subscriptions and providing ROI-focused recommendations for future growth experiments.
Data Scientist
Zomma Protocol | Taipei, Taiwan
- Drove product and market feasibility analysis for DeFi derivatives by producing 10+ research reports based on macroeconomic trends, industry cycles, and quantitative indicators, directly influencing product strategy and business direction.
- Built and automated a scalable data infrastructure by integrating Python with the Deribit API, ingesting and preprocessing 50K+ financial records weekly into MySQL, reducing manual workload by 60% and enabling reliable downstream analytics.
- Designed and delivered a two-stage option pricing system (Black-Scholes + Heston calibration + Monte Carlo simulation) as a core product capability, improving real-time pricing accuracy and increasing trading volume by 20% and market liquidity by 15%.
- Developed an automated dynamic hedging system based on AMM principles, ensuring delta-neutral positioning and effective vega risk control, reducing hedging costs by 12% and increasing platform trading liquidity by 10%, leading to successful production deployment.
Research Assistant
Dept. of Quant. Finance, National Tsing Hua University | Hsinchu, Taiwan
- Reproduced and validated 10+ quantitative financial models in Python by translating top-tier academic research into reproducible implementations, strengthening understanding of modern FinTech trends and production-level financial modeling practices.
- Designed an innovative premium pricing formula for DeFi derivatives to minimize bid-ask spread and improve transaction efficiency, theoretically increasing market liquidity and receiving positive validation from professors and senior analysts at leading securities firms.
- Led weekly academic conferences for 500+ researchers and investors, showcasing cutting-edge FinTech innovations, fostering cross-disciplinary collaboration, and increasing external visibility and recognition of the research team.
Data Scientist
Fubon Financial Holding Co., Ltd. | Taipei, Taiwan
- Designed and improved data quality pipelines by applying over-sampling and SMOTE to address class imbalance, and led feature engineering and EDA on large-scale credit card transaction data, boosting model performance by 25%.
- Built and delivered ML-driven customer scoring models (Random Forest, XGBoost) using 2,000+ customer features to predict repurchase intensity, enabling targeted marketing strategies and driving a 13% increase in repurchase rates.
- Defined model optimization and validation standards by applying PCA, WOE, K-fold cross validation, and resampling techniques, improving predictive accuracy from 60% to 83.89% and achieving an AUC of 0.68.
- Translated model outputs into business-ready insights using SHAP-based interpretability and visualization, aligning technical and non-technical stakeholders and driving adoption of model-driven decision making in production workflows.
EDUCATION
University of Chicago
M.S. in Statistics (STEM)
National Taiwan University
B.S. in Mathematics
Minor in Economics (Business Analytics Program)Relevant Coursework
AI, Data Science & Statistics- Machine Learning, Deep Learning, Natural Language Processing
- Bayesian Modeling, Causal Inference, A/B Testing
- Time Series Analysis & Forecasting
- Generalized Linear Models, Multivariate Statistical Analysis
- Big Data Systems, Statistical Inference
- Probability Theory, Mathematical Analysis, Linear Algebra
- Numerical Analysis, Differential Equations
- Mathematical Finance, Cryptography
- Microeconomics, Macroeconomics
- Econometrics, Applied Economic Modeling
PROJECTS
Daily Digest - AI News Summarizer [GitHub] [Live Demo]
AI Product Design / LLM App / Full-Stack Prototyping / Python / HTML / API Integration / Deployment
- Designed and shipped a lightweight AI news dashboard that aggregates Google News or curated RSS sources (BBC, CNN, NYT, Guardian, WSJ, etc.) and generates “Today’s Key Points” with bilingual summaries (English + optional Chinese translation) powered by fast Groq LLMs.
- Built a flexible LLM backend with an automatic provider switch (Groq / OpenAI / local Ollama), enabling rapid iteration across models and cost/performance tradeoffs through environment-based configuration.
- Product-designed a modern user experience with responsive UI, dark mode, loading skeleton screens, and city-level geolocation detection to personalize news context (e.g., “Chicago, United States”).
- Productionized the prototype with clear project structure, dependency management, and deployment readiness (Dockerfile + Procfile + start scripts), making the app easy to run locally and deploy to platforms like Railway/Render.
Diffusion Model for MNIST Data
Python / Diffusion Model / Deep Learning / PyTorch / Hyperparameters Tuning
- Implemented Denoising Diffusion Probabilistic Models paper, employing UNet architecture to predict conditional mean and error term, achieving clearer image synthesis with the error term model.
- Demonstrated model differences in output, sampling process, and loss function; Model size: 4.3 MB; Training time: 20 secs per epoch.
Optimizing Sentiment Analysis: An Alternative Way to Revisit SESTM Algorithm
Python / MatLab / Big Data / Sentiment Analysis / Text Mining / Hyperparameters Tuning
- Designed and optimized a large-scale NLP pipeline based on the SESTM methodology to extract sentiment signals from 25M+ Dow Jones news articles and predict asset returns, transforming academic text-mining research into a production-level financial signal system.
- Established model benchmarking and optimization standards through systematic hyperparameter tuning, achieving an annualized Sharpe ratio of 5.02 vs. 4.21 baseline, validating both the statistical robustness and real-world trading value of the AI-driven signal.
Big Data - Loan Default Analysis
R / Big Data / Machine Learning / Credit Risk Modeling / PCA / LASSO / Clustering / Product-Oriented Analytics
- Designed an end-to-end credit risk analytics pipeline on 30K+ credit card customer records, performing data extraction, EDA, and feature engineering to support data-driven loan decision frameworks.
- Built and evaluated predictive models (Logistic Regression, Random Forest) using LASSO and PCA for feature selection, achieving 82.09% accuracy and translating model outputs into actionable risk indicators.
- Applied clustering techniques (K-Means, Hierarchical Clustering) to define customer risk segments and uncover key behavioral patterns, strengthening segmentation strategies for credit and marketing products.
Automatic Market Making(AMM) Research Project
AMM / DeFi / Liquidity Mining / Market Design / On-chain Simulation / Product-Oriented Financial Systems
- Designed and evaluated multiple Automatic Market Making (AMM) mechanisms and liquidity mining models, translating theoretical market design into system-level architectures for decentralized trading platforms.
- Built on-chain data simulations to model market-making behavior and liquidity dynamics, validating mechanism performance under different market conditions and usage scenarios.
- Applied dynamic equilibrium models to decentralized finance systems to inform pricing stability, liquidity efficiency, and protocol design decisions.
Big Data Marketing Analysis
SPSS / Big Data / RFM Analysis / CAI & CRI / Customer Segmentation / Recommendation Systems / Product-Oriented Analytics
- Designed a data-driven marketing analytics system using 4,481 supermarket transaction records (2009–2010), applying RFM, CAI, CRI, basket analysis, and clustering to translate raw transaction data into actionable customer insights.
- Built a customer scoring and segmentation framework that quantified purchasing behavior and enabled targeted marketing, recommendation strategies, and personalized customer engagement.
Text Mining DNN Model Analysis
Python / NLP / Deep Learning / PyTorch / BERT / Time Series Forecasting / AI System Design
- Designed an end-to-end NLP pipeline by combining TF-IDF, Word2Vec, and BERT embeddings to transform large-scale unstructured text into high-quality numerical representations for downstream prediction tasks.
- Built and optimized a deep neural network (DNN) forecasting system in PyTorch to predict Australia’s 30-year government bond yield using 1M+ news headlines over 17 years, achieving 95% accuracy and 0.005 MSE through systematic hyperparameter tuning.
Telecom DCB Service Optimization
R / Regression Modeling / GLM / RFM Analysis / Customer Segmentation / Product Optimization
- Designed and built a data-driven optimization system for a telecom Direct Carrier Billing (DCB) service by developing a robust Generalized Linear Model (GLM) in R to identify key drivers of user adoption and transaction performance.
- Translated model insights into actionable marketing and product strategies, defining customer segments and targeting approaches that supported service growth and improved monetization efficiency.
Specific Application of Copula Model, Loss Distribution Approach model (LDA)
- Presented how dependence between risks is important for a fair quantification of the economic capital and introduced the use of copulas to model this dependence.
TECHNICAL STRENGTHS
- AI & Machine Learning: Machine Learning, NLP, Large Language Models (LLMs), Neural Networks, Model Evaluation, Hyperparameter Tuning
- Programming & Data: Python, SQL, R, PySpark
- ML Frameworks: PyTorch, TensorFlow, Scikit-Learn
- Data Systems & Engineering: ETL Pipelines, API Integration, GCP (BigQuery), MySQL, Production Data Systems
- Statistics & Experimentation: Time Series, Regression, A/B Testing, Bayesian Modeling, Causal Inference
- Visualization & Analytics: Looker, Tableau, PowerBI, MS Excel
- Product & Collaboration Tools: Git, Jupyter, Figma (Workflow & System Design), Excel VBA
- Domain: FinTech, CoreBanking, DeFi, Blockchain
EXTRA CIRRUCULARS
- Completed mandatory military service in Taiwan (Spring 2022).
- Participated in the Taiwan FinTech Forum (2021), presenting an On-Chain Fund application concept in campus fintech innovation showcases.
- Attended Meichu Hackathon (2019), designing and prototyping an automatic indoor parking positioning system, strengthening problem-solving and rapid prototyping skills.
- Contributed to National Taiwan University Azalea Festival (2018) by supporting departmental exhibitions and outreach activities.
- Participated in NTU Math Camp (2018), collaborating in technical workshops and peer learning activities.
AWARDS & CERTIFICATIONS
- Product Discovery Certification - Pendo.io
- Career Essentials in Generative AI - Microsoft & LinkedIn
- INSIDE LVMH Certificate - LVMH
- Google Analytics Individual Qualification (GAIQ) - Google
- 1st Place in Mathematics, Regional Science Fair & National Best Local Teaching Material Award, National Science Fair