Data Scientist

Xingzi Xu adijc4@r.postjobfree.com j 802-***-**** j github.com/xingzix j linkedin.com/in/xingzi-xu

Palo Alto, CA, 94305

Education

•

Stanford University Sep 2019 - June 2021

M.S. Management Science and Engineering - Operations and Analytics Track; GPA: 3.80 Stanford, CA

Algorithms, Machine Learning, Deep Learning, Causal Inference, Optimization, Stochastic Modeling, Product Management, Public Speaking, Natural Language Processing (In progress)

•

Middlebury College Sep 2015 - May 2019

B.A. Mathematics & History of Art; GPA: 3.76; UWC Davis Scholarship Middlebury, VT Work Experience

•

Upturn CreditjBBVA USA June 2020 - Present

Data Scientist Intern San Francisco, CA

Created Quicksight dashboards to identify customer pain points in making disputes; results helped prioritize engineering projects and led to 100 more ve-star app ratings on App Store after deployment in a two-week sprint.

Built a Random Forest model to predict dispute outcome with 82% accuracy while maintaining 60% recall; con rmed the feasibility of dispute reminder feature.

Rebuilt ETLs to capture over 500,000 internal status of dispute processes from MongoDB to PostgreSQL and constructed a Markov Chain Diagram to understand unhealthy disputes processes.

Selected features and identi ed reasons of reoccurring errors in credit reports by handling duplicated rows, outliers, and missing values across database with advanced SQL.

•

Kiva Social Good Project (Partnering with Kiva): March 2020 - June 2020 Data Visualization Analyst Stanford, CA

Analyzed the relationship between gender and funding rate by performing hypothesis testing on borrowers’ features.

Used regression to identify posting recency as the most signi cant feature for days to funding, which results in potential country bias on the platform.

Visualized statistical analyses on gender bias and country bias on the Kiva platform; provided recommendations to improve platform justness to Kiva stakeholders.

•

Stanford SeedjStanford Business Nov 2019 - May 2020 Program Management Assistant Stanford, CA

De ned key metrics, conducted cohort analysis, and lifted engagement rate by 30% by conducting controlled experiments on survey data from early-stage Embark entrepreneurship program.

Iterated and launched \high touch" Embark platform by addressing international user pain points based on experiments results. Maintained growth rate and doubled online learning market during Covid.

Communicated with cross-functional teams in 20 countries across India, West and East Africa. Used an Excel macro to increase o ce e ciency, resulting in 10 fewer hours of labor per week for the local program managers.

•

Solar Technology Advisors Feb 2018 - May 2018

Project Assistant Madrid, Spain

Built interactive spider charts for energy price and policy comparison on solar technologies in 13 countries in MENA.

Revised the CSH (concentrated solar heat) nal report in MENA by better articulating the technical reasoning. Projects & Competition

• The Data Open - Citadel Fall 2020 Link: Developed a network dynamics model describing the collaborations between directors and actors. Identi ed statistically signi cant features such as consumer preferences with t-test and used the social network to infer the prestige of individuals in the movie industry over time.

• E-commerce Recommendation System Link: Built an e-commerce recommendation system using collaborative ltering with over 3 million customers’ purchase behavior data. Improved F1 score by 15% using neighbor-based method with TF-IDF weighted matrix.

Skills

• Programming: Python (Numpy, Scipy, Pandas, scikit-learn, seaborn, Tensor ow), R, SQL, C++

• Tools: MongoDB, PostgreSQL, Quicksight, Tableau, Matlab

• Analysis and Modeling: A/B Testing, ETL, Regression, SVM, Tree-based models, Clustering, Simulation

• Languages: Native uency in English & Chinese, Advanced Spanish

Contact this candidate