Xiaofei Li Email & Phone Number

Boston, Ma, Us

• Led data science project portfolios for Auto Physical Damage (APD) claims organization and collaborated with business stakeholders to realize model opportunities and drive prioritization decisions • Led collaboration with MIT to deliver tools that provides interpretability into computer vision models, which build foundation for trust-AI program that enables audits of black-box models• Led collaboration with external vendor like Google and CCC to validate their solutions and made recommendation to senior leaders on vendor decisions• Led data scientists to developed and deployed suites of models that enable Complex claim identification throughout life cycle of casualty claims

Boston, Ma, Us

• Led a project that collaborate among data scientists, engineers, and data stewards to deliver first-ever computer vision models in personal line claims to support claims segmentation• Re-designed and rebuilt Subrogation model for auto line that increased the agreement rate from 30% to 70% and achieved over 70M+ annual subrogation recovery• Led and developed reimagination of Property segmentation model that enabled better channeling accuracy, randomization test, real time threshold update which led to $2.5M annualized incremental savings• Mentored data scientist interns to explore complex Natural Language Processing (NLP) models that using free-form claim notes

Boston, Ma, Us

• Developed machine learning models to identify claims with coverage concern• Expanded NLP use case in claims through developing text mining tools and topic mining models• Mentored junior data scientists to develop time series model for claim volume forecast

Toronto, Ontario, Ca

• Developed CCAR models for different financial stress testing risk management (commercial loans, retail deposits balance, credit card risk, etc…)• Adopted the idea of random forest in variable selection and co-developed a feature selection program in SAS to reduce manual effort• Designed and contributed to a standard modeling procedure that including a 40 pages modeling guidance and over 50 SAS macros (including back-test, sensitivity analysis, stability analysis), which reduces 80% of model development time

Athens, Ga, Us

• Conduct empirical research on issues related to financial institutions (commercial banks and agricultural banks), international trade conflicts; • Compose academic papers based on research and present findings at academic conference, including SAEA, AAEA annual meetings; • Process data and develop regression models for empirical analysis using SAS, R, Matlab and Stata, including time series and cross-sectional methods, panel data methods, generalized linear models; • Developed graphics (figures, maps, tables) for data collected for studies on the Farm Credit System;• Assisted in the implementation of a farm lenders' survey.

Athens, Ga, Us

• Provide statistical consultation to clients from various backgrounds;• Lead independent statistical consulting meeting with clients;• Provided assistance for client including data manipulation, statistical modeling and report writing.

Kansas City, Mo, Us

• Worked closely with bank’s economists to provided research support for branch's regional and rural economic issues, and its publications; • Developed tools for visualizing large financial data sets to identify research topics• Presented research findings to branch AVP on a weekly basis to support the identification of economic issues; • Applied a quantile regression technique to analyze the changes in variability of Agricultural Bank's Return on Assets based on a panel data set containing over 900,000 observations; • Developed and documented data collection and analysis processes.

Doctor Of Philosophy (Phd), Agricultural And Applied Economics

Master Of Science (M.S.), Statistics

Master Of Science (M.S.), Agricultural And Applied Economics

Bachelor Of Arts (B.A.), Economics