Arnon Promjun
Data Scientist
Bangkok
Summary
Data professional with hands-on experience in credit risk modeling, machine learning, and data engineering, leveraging Python and SQL to build end-to-end data solutions. Demonstrated ability to develop predictive models, automate data workflows, and generate actionable insights for risk and lending decisions. Seeking a Data Scientist role to deliver scalable and impactful machine learning solutions.
Overview
2
2
years of professional experience
Work History
Credit Risk Model & Analytics
KASIKORN BANK
01.2024 - Current
- Developed automation solutions for RAROC analysis using Python and Excel VBA, transforming unstructured data into structured datasets and enabling scalable performance evaluation, reducing processing time by ~80% and improving data reliability
- Built automated data workflows for credit cost analysis, standardizing data inputs and generating structured outputs to support consistent monitoring and downstream risk modeling
- Analyzed large-scale transactional data using SQL on Databricks, developing behavioral features and anomaly detection logic (e.g., abnormal transfers, blacklist associations, geospatial inconsistencies, regional anomalies) for fraud detection; automated execution via scheduled workflows and integrated outputs into Tableau dashboards for monitoring
- Analyzed EDC transaction data using SQL to evaluate merchant performance and spending behavior, identifying high-potential merchants for loan offerings and estimating appropriate credit limits to support lending decisions
- Designed and implemented an automated web scraping solution using Playwright to collect court case data, enabling one-click data retrieval and transforming unstructured legal records into structured datasets to support early warning analysis for credit risk and delinquency detection
- Designed and developed bank statement data extraction solutions, evolving from rule-based parsing using Python (pdfplumber) to AI-assisted OCR for image-based documents, enabling scalable and automated feature extraction for downstream analytics
- Developed Exposure at Default (EAD) models within the ECL framework, estimating Credit Conversion Factors (CCF) using 12-month forward exposure at the account level and comparing methodologies between full portfolio and defaulted accounts to refine model assumptions
- Developed forward-looking Probability of Default (PD) models using a Vasicek framework with macroeconomic variables (MEV), generating scenario-based PDs and calibrating weights to adjust baseline PDs for forward-looking credit risk assessment
Data Scientist
FREELANCE - Office of the Consumer Protection Board (OCPB)
12.2021 - 10.2022
- Developed a Multi-Naïve Bayes text classification model in Python to automatically categorize citizen complaints and correct label inconsistencies arising from a multi-level category selection process
- Packaged the model into a deployment-ready module, enabling real-time category validation and suggestion for integration with the complaint web system, reducing misclassification, eliminating redundant processing, and improving case handling efficiency
Education
Master of Science - Statistic, Data Science
CHULALONGKORN UNIVERSITY
Bangkok, Thailand06-2025
Bachelor of Engineering - Industrial Engineering
CHULALONGKORN UNIVERSITY
Bangkok, Thailand05-2021
High School - Sciences and Mathematics
MAHIDOL WITTAYANUSORN SCHOOL
Nakhon Pathom, Thailand01-2016
Skills
Programming: Python, SQL
Data & ML: Pandas, Scikit-learn, Machine Learning, NLP, Feature Engineering
Data Engineering: Web Scraping (Playwright), OCR, Data Extraction (pdfplumber), Automation
Tools & Platforms: Databricks, Power BI, Altair, Excel (VBA)
Domain: Credit Risk Modeling (ECL, EAD, PD, RAROC, CRR), Credit Cost & Lending Analytics
Timeline
Credit Risk Model & Analytics
KASIKORN BANK
01.2024 - Current
Data Scientist
FREELANCE - Office of the Consumer Protection Board (OCPB)
12.2021 - 10.2022
Bachelor of Engineering - Industrial Engineering
CHULALONGKORN UNIVERSITY
High School - Sciences and Mathematics
MAHIDOL WITTAYANUSORN SCHOOL
Master of Science - Statistic, Data Science
CHULALONGKORN UNIVERSITY
University Projects
SENIOR PROJECT – Twitter Sentiment Analysis on Tourism During COVID-19
- Developed and evaluated classification models (SVM, Decision Tree, Random Forest) on tourism-related Twitter data during COVID-19, selecting SVM as the optimal model
- Applied NLP techniques for text preprocessing and feature engineering (e.g., tokenization, stopword removal, vectorization) to transform unstructured tweets into structured inputs
- Used K-Fold cross-validation and Grid Search for robust model evaluation and hyperparameter optimization
- Visualized sentiment trends to derive insights on public perception and tourism impact during the pandemic
THESIS – Active Learning for Text Classification
- Utilized the same dataset as the sentiment analysis project with a different modeling objective, focusing on data efficiency rather than predictive performance
- Implemented active learning strategies with reinforcement learning concepts on tourism-related Twitter data during COVID-19 to optimize labeling efficiency under limited labeled data
- Compared Random, Greedy, and Thompson Sampling (with Laplace Approximation), demonstrating that Thompson Sampling outperformed other methods in balancing performance and labeling cost
Publications
- Sontayasara, T., Jariyapongpaiboon, S., Promjun, A., Seelpipat, N., Saengtabtim, K., Tang, J., Leelawat, N., Twitter sentiment analysis of Bangkok tourism during COVID-19 pandemic using support vector machine algorithm, Journal of Disaster Research, 16, 1, 24-30, 2021
- Leelawat, N., Jariyapongpaiboon, S., Promjun, A., Boonyarak, S., Saengtabtim, K., Laosunthara, A., Tang, J., Twitter data sentiment analysis of tourism in Thailand during the COVID-19 pandemic using machine learning, Heliyon, 8, 10, e10894, 2022