Yingxiang Zhang's Email Addresses & Phone Numbers

📖 Summary

Director, Quantitative Risk @ UBS Greater New York City AreaAssociate Director @ UBS From January 2018 to February 2020 (2 years 2 months) Greater New York City AreaQuantitative Analyst, Assistant Vice President @ Bank of the West Credit Risk Modeling:* Developed Auto PD and LGD models from scratch, and re-built CRE PD and LGD model by including external data (Trepp CMBS performance data) for DFAST/CCAR and risk management processes.* Built an origination scorecard for Equipment Finance Division’s upcoming Canadian portfolio. This scorecard model is designed to help in the decision-making process at origination stages for equipment leases in Canada.* Assessed quality of model outputs trough back testing against realized outcomes, benchmarking against alternative models, sensitivity tests around assumptions and limitations, and other relevant tests.* Implemented the CCAR models to create nine-quarter prediction under internal and supervisory scenarios, and explained the model performance by looking at the trend of key risk drivers.* Monitored the model performance periodically following a three-tiered escalation plan, and quantified the shifts in population dynamics using Population Stability Index (PSI).* Assisted in completing the Bank's Allowance for Loan and Lease Loss (ALLL) reserve allocation and IFRS9 expected losses to senior and executive management.Anti-Money Laundering:* Served as a developer responsible for the creation and optimization of KYA (Know Your Account) models and machine learning monitoring tools in support of financial investigative efforts.* Retrieved the transaction data from SAM8 system, and aggregated the transaction data into daily records by four transaction types - Wire/ACH/Cash/MI.* Built a Dynamic Bayesian Network model to detect suspicious patterns by identifying anomalies in sequences of transactions. Benchmarked the Dynamic Bayesian Network against alternative machine learning algorithms, such as Naïve Bayesian, Markov chain and time-series clustering.* Tuned the model parameters by maximizing the AUC and related statistics. Defined the methodology/approach in a clear and concise manner to support the model implementation efforts. From June 2016 to January 2018 (1 year 8 months) San Francisco Bay AreaSummer Intern, Loss Forecasting @ Ford Motor Credit Company • Manipulated and reorganized large data sets according to delinquent types by using SAS (SQL), and developed a time-series model that improved our ability to forecast recoveries over the short term.• Preformed model validation to CPE (a vintage curve model used by Ford Credit to forecast PD & credit losses), and conducted scenario analysis to test sensitivities to assumption and scope the range of outcomes.• Corrected an invalid assumption with multinomial logistic regression and improved the forecasting accuracy of CPE. Coded in VBA to format the model outputs and generate recovery forecasting reports.• Presented the models to COO and leadership team, and won award of excellent for Executive Challenge Program. From June 2015 to August 2015 (3 months) Greater Detroit AreaEquity Research Intern @ Elle Investments Inc • Analyzed NASDAQ-listed biotech companies with respect to the market and financial condition.• Estimated revenues and cash flows under different scenarios, and preformed valuation by using DCF.• Made recommendations on possible positions and built VBA macros to monitor profit and loss (P&L) on a daily basis. From December 2014 to April 2015 (5 months) Houston, Texas Area