Yipeng Wang

Explore, process, and load data sets into the system to develop signals for systematic quantitative strategies; Develop market-neutral and medium-frequency alphas that predict future return; Combine alphas to make final quantitative strategies that can be traded in the market by machine learning and statistical methods; Involved markets include equity markets in US and China and futures markets in US.

• Through the data of user's action on the Grab APP and user’s historical activities, Grab predicts fraudulent behaviours (chargeback fraud, etc.) to intercept fraudulent bookings when user is ordering; • Leverage explainable AI (e.g. SHAP) to explain this black-box model's prediction, to inform business team or blocked users the reason of block.

• Migrated an ETL pipeline to DBS's Spark cluster on AWS, which is to construct a capital flow network of DBS's clients, using Alluxio, Eclipse Theia, and Pyspark on TB of data; • Analysed the data pipeline to seek and optimize its performance bottlenecks, including substituting TF-IDF with LSH to reduce the time of company name matching.

Biostatistics research