- Design machine learning solutions with time series analysis on over 1 billion customers’ credit report data- Helped banks to minimize risks through a better estimation of time-to-default and a better prediction on cash flows- Wrote SQL codes to retrieve requested data from database and prepared the data for modeling purposes- Constructed Wavelet Transformation on 5 features by generating 10 scaling levels for each- Built Random Forest models using python to predict the probablity of default up until different months- Performed Survival Analysis on 24 nested time windows to predict time-to-default

- Analyzed the linguistic strategy that President Trump used on Twitter to reinforce his governance and influence- Classified his tweets into different linguistic types and assigned scores based on their frequency in each tweet- Used statistical techniques for hypothesis testing to validate interpretations: built regression models to forecast how the linguistic styles (negative emotions, informal languages) impact the number of retweets, replies, and likes- Offered interesting insights into how language use may serve as a political strategy to reinforce governance

Marketed financial products to clients, including mutual funds, stocks, and bonds.