Zexi (Jersey) Liu Email & Phone Number

Amsterdam, Nh, Nl

Netherlands

Amsterdam, North Holland, Netherlands

• Coordinated with stakeholders and data engineers to fix data quality issues.• Applied analysis on data lineage and Airflow DAGs to improve data quality.• Collaborated closely with data engineers to develop data models and ETL pipelines.• Employed SQL and Excel to construct a cohort analysis, which facilitated a new viewpointand provided valuable insights into the lifetime value of the membership program.• Wrote readable and maintainable SQL as a template for future A/B… Show more • Coordinated with stakeholders and data engineers to fix data quality issues.• Applied analysis on data lineage and Airflow DAGs to improve data quality.• Collaborated closely with data engineers to develop data models and ETL pipelines.• Employed SQL and Excel to construct a cohort analysis, which facilitated a new viewpointand provided valuable insights into the lifetime value of the membership program.• Wrote readable and maintainable SQL as a template for future A/B test analyses.• Constructed a pySpark and Python pipeline to perform significance testing, resulting ina workflow that was more than 10 times faster than before.• Coordinated with data scientists to analyze A/B test of various models, and enhanced theTableau dashboard to visualize the results of different experiments in a more clear andvisually appealing manner. Show less

Amsterdam, North Holland, Netherlands

• In order to train machine learning models, used BigQuery and Firestore for dataextraction.• Employed Python for data transformation and cleaning. Coordinated with stakeholdersto gain insights for feature engineering.• Built a text classification model with a f1 score of 0.9, and a duplicate logo detectionmodel with a f1 score of 0.88. Then built and deployed APIs for these models via FastAPI.• Built multiple web scrapers. Then deployed and maintained Airflow for these… Show more • In order to train machine learning models, used BigQuery and Firestore for dataextraction.• Employed Python for data transformation and cleaning. Coordinated with stakeholdersto gain insights for feature engineering.• Built a text classification model with a f1 score of 0.9, and a duplicate logo detectionmodel with a f1 score of 0.88. Then built and deployed APIs for these models via FastAPI.• Built multiple web scrapers. Then deployed and maintained Airflow for these scrapers.• Gained experience working with CI/CD and Docker in doing above projects. Show less

Beijing, China

- Improved the weekly retention of recipe generators from 18% to 30%- Conducted multiple A/B tests to determine the best position to show new recipes in search engine results, achieved 2000%+ growth to their number of views- Used SQL, Excel, and Tableau to conduct multiple analysis and provide insight - In charge of managing ad-hoc data analysis tasks

Master'S Degree, Data Science For Decision Making

Affective Video Summarization: - Determine the most appropriate dataset and model for the emotion recognition part of our project.

Master'S Degree, Human Decision Science

Bachelor'S Degree, E-Commerce/Electronic Commerce