Data Scientist

• Feb 2019 - Dec 2020

Grader

• 2018 - Dec 2018

Grade each project(from methods, code reproducibility and presentation, etc) and answer questions students have. Introduce methods I have used in previous projects.

Teaching Assistant

• Feb 2018 - May 2018

• Held office hour to answer questions for course4282 linear regression model and time series. • Corrected the midterm and final exams paper and provide solutions to each homework and exam. • Taught midterm exam and prepare a review for linear regression to 45 people's class.

Intern, Data Scientist

• May 2018 - Aug 2018

• Designed Cardinal Operation’ s product StockGo, an automated model building software. Used Python to autonomously extract relevant features (such as holidays, promotions, average sales, etc) from time series data, then build time series, machine learning, and LSTM models. StockGo predicts a time period’s sales according to customer’s time granularity requirement and methods for prediction (the product will extract relevant features and tune model parameters by itself). • Used Python to… Show more • Designed Cardinal Operation’ s product StockGo, an automated model building software. Used Python to autonomously extract relevant features (such as holidays, promotions, average sales, etc) from time series data, then build time series, machine learning, and LSTM models. StockGo predicts a time period’s sales according to customer’s time granularity requirement and methods for prediction (the product will extract relevant features and tune model parameters by itself). • Used Python to extract customer features (such as mean and variance of billing amount, packaging fee, cargo weight and volume, etc.) from Deppon Express's 8 million domestic waybill data and 1000 overseas waybill data. Used overseas data to identify domestic clients with potential overseas demand. Iteratively identified 2000 candidates using cosine distance with randomly extracted features. Took intersection of 2000 candidate sets as a prediction for domestic clients with potential overseas demand. Show less • Designed Cardinal Operation’ s product StockGo, an automated model building software. Used Python to autonomously extract relevant features (such as holidays, promotions, average sales, etc) from time series data, then build time series, machine learning, and LSTM models. StockGo predicts a time period’s sales according to customer’s time granularity requirement and methods for prediction (the product will extract relevant features and tune model parameters by itself). • Used Python to… Show more • Designed Cardinal Operation’ s product StockGo, an automated model building software. Used Python to autonomously extract relevant features (such as holidays, promotions, average sales, etc) from time series data, then build time series, machine learning, and LSTM models. StockGo predicts a time period’s sales according to customer’s time granularity requirement and methods for prediction (the product will extract relevant features and tune model parameters by itself). • Used Python to extract customer features (such as mean and variance of billing amount, packaging fee, cargo weight and volume, etc.) from Deppon Express's 8 million domestic waybill data and 1000 overseas waybill data. Used overseas data to identify domestic clients with potential overseas demand. Iteratively identified 2000 candidates using cosine distance with randomly extracted features. Took intersection of 2000 candidate sets as a prediction for domestic clients with potential overseas demand. Show less

Intern, Counselor, Advisory-consulting-experience center-data analytics team

• Mar 2017 - May 2017

• Collected and organized company profiles of Bank of China clients and their relations with Bank of China. • Analyzed Bank of China data to set data-cleaning guidelines and produce data-quality reports. • Imported and cleaned data using SAS based on data-cleaning guidelines, processed and transformed data according to the demand of models and data quality report. • Collected and organized company profiles of Bank of China clients and their relations with Bank of China. • Analyzed Bank of China data to set data-cleaning guidelines and produce data-quality reports. • Imported and cleaned data using SAS based on data-cleaning guidelines, processed and transformed data according to the demand of models and data quality report.