Improved a PCA model to forecast mid-price and yield for U.S. Treasury instruments. Providing support for market-making and hedging strategies for U.S. Treasury and European Government Bonds.

Developed a NLP algorithm for automatic table structure identification. Built a deep-learning model for automatic table content interpretation. Delivered 3 patents. “Automatic Table Understanding” poster was accepted by 2022 INFORMS.

Built a Customer Relationship Management (CRM) system based on Enterprise Resource Planning (ERP) system to gather departments customer information