Utilized NLP techniques to analyze earnings reports, creating a LLM model for enhanced financial forecasting accuracy. Fine-tuned BERT model focusing on sentiment analysis to enhance precision in data interpretation and predictive analytics.

Analyzed technological capabilities in key sectors, providing insights for strategic product positioning. Performed competitive analysis of database products, benchmarking and strategizing for market advantage. Outlined key risks in cross-border M&A, integrating market research into a comprehensive product risk management strategy.

Completed a development proposal for a city pilot in aluminum, agricultural, and manufacturing industries. Assisted clients in identifying eligible inventions and patents for tax exemptions and incentives. Created technology agreement contract templates and transfer pricing documents based on historical records

Use R to implement the “quantitative bond timing and duration rotation” model. Take fundamental factors and bond valuation factors to predict the yield of 10-year China Treasury bonds. Reduce the maximum backtest by 7%.Track credit bond issuance and bidding in the primary market and the secondary market yield trend; analyze the trend of credit bond's maturity and rating spreads, and combine market sentiment and macro policies to give investment recommendations.Collect, analyze and share the sell-side views in the morning meeting, including the latest fixed-income industry news, macroeconomic trends, central bank monetary policy, research reports on specific convertible bonds, etc.

Use NumPy and pandas of Python to analyze the real estate market detailed, calculating 20 kinds of data including the leverage ratio and growth potential. Build multiply time series models, observe the abnormal turning points and present case analysis.Use Wind's Api in R to tabulate the credit rating reports of different companies with their stock and bonds, use ggplot to draw the histograms of ROE and profit to find the outliers, identify the reasons of the outliers.Create a word rating system using SQL to judge the positive and negative sentiment about company news. Use weighted algorithm to rate the word with sentiment points, and judge the positive/negative effects on credit ratings.

Restructured companies' big data using SQL, clean the dirty data and build the initial connection between companies' database and Oracle Cloud.Upgraded the 2020 document of Oracle Cloud, including coding, testing, and document editing. Present the advantages of Oracle Cloud including convenience in SAAS and creative tools in PAAS.Constructed data visualization by Endeca and ODV. Find the most appropriate visualization method of different data structures.