Ph.D. candidate in Applied Mathematics and Statistics with Time-inhomogeneous Lévy Processes research area. My current research topic is the Lévy processes applications in the TSIR models and bond derivatives pricing.
My Ph.D. advisor is James Glimm, who is the former president of the American Mathematical Society and a recipient of Heineman Prize, Leroy P. Steele Prize and National Medal of Science.
Actively look for Quantitative Reseach and Data Science position in finance and technology industry.
• Monitored $67 billion multi-asset allocation and conducted optimal portfolio analysis using different theories, e.g. extreme value theory - Pareto distribution, mean-variance
• Simulated tracking errors of global fixed-income portfolios by modifying parallel/non-parallel duration shifts on each bucket and applied different weight sampling methods to find distributions of tracking errors
• Evaluated external managers’ performance using Stochastic Dominance, Regime Shift and Monte Carlo simulations for forecasting and VaR/ES calculation
• Performed risk mapping and simulation of $15.3 billion global private equities and real assets
• Lectured class; prepared LaTeX class notes; constructed and graded assignments and tests (AMS 318 Financial Mathematics and AMS 320 Quantitative Finance)
• Applied data modeling and statistical learning methods to market prediction and systematic trading
• Calibrated term structure of interest rate models in use of interest rate risk management
• Managed the structuring, forming and operations of private investment funds, including Private Equity Funds, Real Estate Funds, Fund of Funds
• Implemented the solution of fund administration and banking for Delaware funds and Cayman funds
• Conducted due diligence on the underlying assets and contributed to product development
Advisor: James Glimm
Research Area: Time-inhomogeneous Lévy Processes
• Implement one-vs-all logistic regression and neural networks to recognize hand-written digits
• Built spam classifier using support vector machines (SVMs) with Gaussian Kernels
• Replaced classical assumption of a linear model for the distribution of a random vector is replaced by the weaker assumption of a model for the copula. Estimated dependence by using Kendall’s τ and performed copula structure analysis on a financial data set and explained differences between copula-based approach and the classical approach