DoorDash for Business

Tech Lead of Loyalty & Membership (Lyft Pink)- Core benefits- Financial framework- Growth- Platform

1. Predictive Model Development: built scenario-based Probability of Default (PD) Forecast model for CECL/CCAR usage (covering over 90% of commercial lending portfolio; approaches used: time series, generalized additive models, hierarchical clustering)2. Quantitative Research: explored and applied industry-leading modeling approaches and performance metrics to enhance the current modeling and monitoring process 3. Outcome Analysis: analyzed model output and performance each cycle and made recommendations to stakeholders

Thesis: Investigated transformation invariance of convolutional neural network (CNN)1. Simulated image data and compared classification accuracies of CNN on data under different types of transformation2. Investigated the function of model intermediate layers in dealing with data transformation3. Coached junior graduate student in data simulation and model optimization for next steps in research

1. Teaching Assistant: Taught ~50 students each semester general chemistry. Responsibilities included fostering class discussions, instructing students on how to conduct experiments, helping students with homework and grading lab reports.2. Research Assistant: Studied the efficiency of a surface‐activated material in extracting biomarkers for cancer detection and optimized the capacity and bio-compatibility of the material.

Developed a software package to visualize the singular value decomposition (SVD) components of data. Programmed and compared different SVD‐based visualization methods systematically. Analyzed an internet traffic data set to explore usage patterns and abnormal features.