Designed and implemented data pipelines and schemas to automate Daily Balance Sheet reporting. Developed and configured Power BI dashboards with auto-refresh capabilities, reducing reporting time by over 90% and significantly improving operational efficiency.
Developed analytical solutions by implementing appropriate statistical tests and machine learning models using Python to verify current deposit runoff factors. The suggested results were reviewed and accepted by the CFO, leading to adjustments in the corresponding metrics based on my findings.
Automated the integration testing process by developing a Python package to parse configuration files, identify SQL typos and syntax errors, and generate log files. This automation reduced the time required for package integration testing by at least 50%.
Developed machine learning models (Random Forest, XGBoost, and Gradient Boosting) to analyze recent bank failures and predict deposit withdrawals for BMO during extreme stress periods. Presented the final results to the entire Treasury team, receiving an extraordinary rating from my manager for the project.
Associate, Wholesale Credit Modelling
MUFG Bank, Ltd.
12.2019 - 12.2021
Developed a wholesale PD scorecard model from inception to completion, addressing imbalanced data issues and integrating both quantitative and qualitative models to calibrate the final credit score. The newly developed model improved accuracy by approximately 20%.
Developed a Python library using BeautifulSoup to parse SEC EDGAR filings and generate DataFrames of financial information for selected companies. The Final package was saved and employed by the team members since December 2020 to improve work efficiency.
Conducted independent reviews of various developed models and generated review reports upon identifying inconsistencies or triggering model alarms.
Researched and evaluated outlier detection methods, including Z-score, ESD Test, Pierce's Criterion, Isolation Forest, Outlier Removal Clustering, and DBSCAN. Selected the most suitable method for the dataset, resulting in a 20% improvement in PD model accuracy.
Credit Risk Data Analyst Co-op
Royal Bank of Canada
Toronto, ON
01.2019 - 04.2019
Applied SQL query to extract and calculate ACL, PCL, ECL and other risk parameters (PD, EAD and LGD) for RBC's commercial portfolio to provide effective on-site measuring and monitoring of material risks for Business Financial Services.
Helped to build a python process to clean and read clients' mortgage information into database
Provided insight and analysis to senior management on risk limit utilization, credit migration and trends.
Supported statistical and quantitative analysis on a comprehensive deep-dive of the Commercial Real Estate Portfolio, identifying the optimal proportion of Real Estate Portfolio