Potchara Pheungthonglor
Data Science Intern · AmeriLife
Los Angeles, CAWork Experience
Data Science Intern
AmeriLife2022 – 2022
? Developed a two-stage hurdle model (classification & regression) for agent sales prediction, evaluated the LightGBM models under various distributions to handle the zero-inflated target, and built ML pipeline in Kedro ? Identified 8 states and 800 counties in the US as potential Medicare Advantage growth markets by clustering analysis, visualized the characteristics for each cluster, and conducted the presentation for the vice president ? Implemented 3 automated data pipelines for scheduled data extraction and analysis from APIs using Kedro
Advanced Data Scientist
KBTG2019 – 2021
Built revolver assets using customer micro-segmentation and look-alike modeling: RF to launch >30 targeted credit card and fast loan campaigns that boost ~$150 million incremental spending of revolvers in 2020 ? Spearheaded data science team in Machine Lending project that won the Best Innovative Company Awards 2019 from the Stock Exchange of Thailand in recognition of the AI-based machine lending service ? Supervised data scientists and interns on 6 marketing analytics projects: product recommender system, customer persona analysis, and propensity to revolve modeling that won KBTG Bootcamp: Best business analytics award ? Provided insights and recommendations on 4 push-pull marketing strategies to the executive leadership team ? Researched and applied state-of-the-art methodologies: reinforcement learning recommendation system, DNN,
Data Scientist
CNN, RNN2017 – 2019
? Developed an automated process encompassing data ingestion, propensity modeling: SVM, RF, and XGB with lift up ~5X from baseline, and ML pipeline deployment that disbursed >$1.2 billion of personal loan/year ? Implemented automated ML pipelines and workflows using PySpark and sklearn, reduced >150 man-days/year ? Created proprietary car appraisal models applying deep learning methods: ensemble image segmentation and classification using YOLO, fast.ai, and PyTorch to predict used car specifications and price ? Programmed web scraping using Selenium to create annotated car dataset of >1 million images automatically
Research Intern
Division of Media and Network Technologies, Hokkaido University2016 – 2016
Performed data analysis and visualization on elderly Japanese patients data to derive insights and created a web application to report health issues to doctors from home via IoT devices ? Built and fully developed a virtual reality game that synced with the Oculus Rift
Teaching Assistant (Algorithm Design, Discrete Mathematics, Computer Programming)
Department of Computer Engineering, Kasetsart University2014 – 2016
Bangkok, Thailand Teaching Assistant (Algorithm Design, Discrete Mathematics, Computer Programming) Jan 2014-May 2016 ? Lectured and designed laboratories, coursework, and exams for > 200 undergraduate students
Education
B.Eng. in Computer Engineering
Kasetsart University2017
B.Eng. in Computer Engineering
M.S. in Computer Science
University of Southern California2023
M.S. in Computer Science