-Improved the accuracy of an existing XGBoost machine learning model by 2% through analyzing 4.5M records of social welfare candidate data-Improved the model's runtime performance by 10% by implementing Light GBM algorithm-Provided an alternative solution for a better model prediction by applying time-series analysis and allocating more statistical weights to unpredicted candidates-Assembled the development server by utilizing Docker container, R-Studio, Oracle R-Distribution, and JupyterLab-Utilized Sybase SQL to query variable data based on specific conditions and modified the database with updated raw data-Visualized data points and outlier boundaries on plots through the ggplot2 package in R-Served as a liaison between clients and internal team to ensure effective data analysis, appropriate model expectation and goal alignment by scheduling weekly meeting and providing meeting summaries

-Performed analysis on property types and midstream industry for potential investment opportunities mainly in condominium, hotel, and new energy industries -Presented key highlights and results of previous projects such as investments in liquified natural gas export company in the US and Public-Private Partnership (P3) railroad in the UK to team members -Prepared meeting minutes for conference calls with investment partners to discuss project schedules, collaterals, administrative agents, and loan agreement adjustments