• Worked on preprocessing training data for Fast Text and Doc2Vec model including parsing HTML, flatten lists, removing stop words, stemming and tokenization

• Analyzed feedbacks on Article Highlight model, provided visualized data summarization for tuning model

• Built debugging tool for Article Highlight project to improve algorithm performance

• Worked on training, deployment, knowledge transfer documentation of Article Highlight model

• Worked on developing Doc2Vec, MLP and random forest classifier for Quick Reply project

• Time series analysis mock-up in Python using statistics library for forecasting on cash flows segments

• Carried out validation on massive data in order to provide clients with good quality cash reports

• Coordinated with Development Team, analyzed on bugs and improved the existing code for the application

• Set up cash forecasting logic depending on historic payment rolls

• Configured and compiled customization for new client’s application in QlikView for ETL process from ERP system or bank statements for visualization and KPI purposes

• Built program solely for Event Study, which checks and analyze the significance of different events’ influence on stock return

• Extracted data from SQL Server, computed statistical summary, visualized data through Matplotlib

• Modeled cumulative abnormal return (CAR) of classified stocks over benchmark within a 3 months’ time window through Numpy, Pandas, Datetime packages implemented in Python using OOP design

• Developed program that auto-generated daily stock investment filter based on events (news, announcements etc.)

• Detected positive and negative influence events out of 17 events, provided outline for reference and future updates

• Worked closely with senior analysts, performed feasibility studies, financial analysis and due diligence

• Conducted top-down analysis by comparing real estate markets in the West, North East and South East region of the US based on population, industrial structure, education and etc. The assembled data was then categorized into an Excel spreadsheet for investment evaluation