Developed responsive web interfaces using TypeScript and Next.js with Ant Design; created an Image Search feature powered by VGG16 for image-based product search deployed via Nest.js; enhanced multilingual data analysis capability using NLP for Thai to broaden applicability in customer reviews; fine-tuned a Bag-of-WWords based XGBoost model for user review spam detection and integrated it into the website.

Data Scientist at Metro Systems Corporation Public Company Limited

December 2, 2024 - October 22, 2025

Collaborated with researchers to develop a fine-tuned XGBoost model with PCA for accurate Above Ground Biomass (AGB) estimation, reducing RMSE by 11%; identified key factors for AGB estimation via Boruta and Recursive Feature Elimination; built an uncertainty model to validate Satellite-LiDAR data as an alternative to field data, achieving ~4x improvement in uncertainty accuracy with 5–10% variance; designed and deployed an AGB estimation platform enabling real-time AGB and uncertainty estimates, reducing processing time from weeks to minutes; partnered with data management to build scalable, reproducible data pipelines covering cleaning, labeling, feature engineering, and transformation; developed a production-ready XGBoost model for product-level churn prediction with recall > 95% via API.

AI Research Intern at National Bio Bank of Thailand

September 30, 2023 - October 22, 2025

Collaborated with researchers in Pathum Thani to support AI-driven projects, contributing to model development and evaluation, including data preprocessing, feature engineering, and exploratory analysis for biomedical datasets.

Software Development Intern at Metro Systems Corporation Public Company Limited

December 1, 2022 - December 1, 2022

Developed responsive web interfaces with TypeScript and Next.js with Ant Design. Created an image search feature powered by VGG16 for image-based product search, deployed via Nest.js. Enhanced the model's capability to analyze multilingual data using NLP for Thai, broadening its applicability and improving overall customer review analysis. Fine-tuned a Bag-of-WWords based model for user review spam detection and integrated it into the website.

Data Scientist at Metro Systems Corporation Public Company Limited

December 2, 2024 - December 2, 2024

Collaborated with business teams to analyze vendor software, providing insights into data solutions. Fine-tuned an XLM-RoBERTa large language model (LLM) for churn classification, achieving improved accuracy and multi-language support. Built scalable data pipelines covering data cleaning, labeling, feature engineering, and transformation. Built a fine-tuned XGBoost model for product-level churn prediction, achieving high recall and deployed as an API.