About Me

I am Cheng, currently based in Tokyo as a Data and AI Consultant at arcbricks, leading cross-border teams to harness the potential of Databricks and data to foster digital transformation.

I am a Master’s graduate in Information Science from Cornell University. Previously, I received my Bachelor’s degree in International Business from National Taiwan University with a minor in Sociology.

My experiences span data analysis, management consulting, brand strategy, and academic research, and I am interested in how to leverage data-driven insights in diverse contexts.

Education

• Master’s in Information Science, Cornell University, 2024 - 2025
• Bachelor’s in International Business, National Taiwan University, 2018 - 2023
• Bachelor’s in Sociology (minor), National Taiwan University, 2018 - 2023
• Business Analytics Program, National Taiwan University, 2018 - 2023

Work Experience

• Data and AI Consultant at arcbricks, Oct. 2025 to present (Tokyo, Japan)
• Consultant Intern (part-time assistant) at beBit TECH, Feb. 2022 - Jun. 2022 (Taipei, Taiwan)
• Strategy Intern at Redpeak, Aug. 2021 - Dec. 2021 (Taipei, Taiwan)

Research Experience

• Research Assistant under the supervision of Professor Hsuan-Wei Lee at Lehigh University, Nov. 2024 - Present
• Research Assistant under the supervision of Professor Nanyi Bi at National Taiwan University, Oct. 2023 - Jul. 2024
• Research Assistant under the supervision of Professor Yen-Hsin Cheng at the Institute of Sociology of Academia Sinica, Oct. 2022 - Dec. 2022

Project Experience

Ecolab Food Safety Audit Question Recommendation System

• Constructed a recommendation system that provides actions for over 600 FDA audit questions for Ecolab, one of the industry leaders in offering water, hygiene, and infection prevention solutions
• Leveraged data parsing and embedding techniques to preprocess data from over five different data sources
• Utilized Databricks to build a scalable Zero-Shot recommendation system that provides recommended actions, reasoning, citations, and sources for each corresponding audit question

Estimation and Visualization of Flood Damage in Taiwan (link)

• Aggregated data from five different souces, compiling over seven million flood observation entries
• Utilized spatial and temporal data to identify 84 unique flood events across Taiwan
• Developed a dynamic Tableau dashboard to visualize estimated flood damage over five years

Customer Segmentation and Business Analysis Report Based on Transaction Data (link)

• Generated a business analysis report from 541,909 entries of transaction data of an online retailer based in England
• Defined four behavioral indexes (RFM Indexes, Activity Index) as features for K-Means Clustering and identified three distinct clusters
• Delivered three levels of next-step business strategy suggestions based on the segmentation results and the overall patterns in the data

Prediction of League of Legends Game Outcome with Supervised Learning Methods (link)

• Predicted the outcome of League of Legends games based on in-game data from a dataset with 28 features across 24,224 entries.
• Engineered supervised learning methods such as Logistic Regression, Linear Discriminant Analysis, Random Forest, and XGBoost, yielding a final prediction accuracy of over 78% across all four models.
• Leveraged hyperparameter tuning methods such as Grid Search and Bayesian Optimization to further optimize model performance.

News Category Classification with Natural Language Processing and Deep Learning (link)

• Categorized the category of news entries based on the headline and description from a dataset with over 200,000 entries, deploying both classical and deep learning machine learning models.
• Implemented Principle Component Analysis to perform dimension reduction and feature transformation.
• Applied Logistic Regression, Support Vector Machine, and Multilayer Perceptron with Bayesian Optimization hyperparameter tuning, achieving a final prediction accuracy of 74% on the best-performing model.

Impact of Xiaohongshu Usage on Young Female’s Body Dissatisfaction (link)

• Analyzed how social media influences females’ body dissatisfaction with Xiaohongshu as an instance under the instruction of Professor Nanyi Bi from the Department of Information Management at National Taiwan University.
• Conducted bottom-up thematic analysis and constructed 9 code groups with over 100 codes across 11 in-depth interviews with MAXQDA.
• Crafted a conceptual model depicting the interplay between social media usage, body dissatisfaction, and body comparison while considering the characteristics of Xiaohongshu’s content and usage patterns.

Software, Skills, and Languages

• Programming Languages

  • Python (excellent), R (excellent), SQL (excellent), C++ (intermediate), VBA (intermediate)

• Skills

  • Supervised Learning Methods: Regression, Discriminant Analysis, Support Vector Machine, K-nearest Neighbors, Multilayer Perceptron, Random forest, XGBoost
  • Unsupervised Learning Methods: Principal Component Analysis, K-means Clustering, Latent Dirichlet Allocation, Latent Class Analysis
  • Generative AI Engineering: Embedding Pipelines and Evaluation, RAG Architecture Design, Prompt Engineering, Evaluation and Experiment Management with MLflow
  • Data Engineering: Data modeling, ETL/ELT, Workflow Orchestration, Medallion Architecture Design, SCD, Quarantining

• Tools

  • Data Platforms: Databricks (Data Modeling, Data Pipeline Engineering, Model Management with MLflow, Feature Store, Dashboards, Feature Store)
  • Project Management: Agile/Waterfall Project Management with Azure DevOps, Jira
  • Version Control: GitHub
  • Business Intelligence: WhereScape RED (ETL and Data Modeling), Tableau (Data Visualization)
  • Statistical Software: STATA, SPSS

• Credentials

  • Databricks Certified Data Analyst Associate (credential link)
  • Databricks Certified Generative AI Engineer Associate (credential link)
  • Databricks Certified Data Engineer Associate (credential link)
  • Databricks Certified Machine Learning Associate (credential link)
  • Databricks Certified Machine Learning Professional (credential link)

• Languagues

  • Mandarin (native), English (native/bilingual proficiency, C2), Japanese (professional working proficiency, JLPT N1)