Experience

Code Willing

• United States
• IT Services and IT Consulting
• 1 - 100 Employee

• Data Product Analyst

  • Mar 2023 - Present

  Data Science Team: Migrate Terabytes level financial data from traditional RDBMS to DataLake Data Science Team: Migrate Terabytes level financial data from traditional RDBMS to DataLake



Rensselaer Polytechnic Institute

• United States
• Higher Education
• 700 & Above Employee

• Data Analyst

  • Oct 2021 - Dec 2022

  • Predicted student retention using XGBoost algorithm on 10-year cohort data from different departments, and provided watch lists for college stakeholders for early interventions, which enhanced student retention rate by 2.8%. • Addressed imbalanced target variable problems with oversampling method, and utilized Bayesian Optimization approach (HyperOpt) to improve model performance by 30% F1 score on average. • Designed a semi-automated data processing and modeling package (Python, Git) to streamline coding tasks for team members. • Developed and migrated dashboards from local Excel Pivot Table to Tableau Online server, which automated the process of monitoring students’ weekly learning status. Show less



Velan Studios, Inc.

• United States
• Computer Games
• 1 - 100 Employee

• Data Scientist (Capstone)

  • Jan 2022 - May 2022

  • Coordinated with a group of 5 as technical lead, initiated and managed data processing and modeling pipeline on Google Colab. • Built predictive models to forecast player satisfactions based on 70+ in-game metrics for Knockout City video game. • Applied Logistic Regression model (Statsmodels) on different player segmentations to understand drivers of player satisfactions. • Generated dynamic dashboards (Tableau) and business reports based on subgroup analysis towards 300,000+ in-game matches. Show less



Tencent

• China
• Software Development
• 700 & Above Employee

• Data Analyst

  • Nov 2020 - Jan 2021

  • Established a user funnel model for WeChat Work Microdocument product, and monitored key metrics such as retention rate and conversion rate to evaluate A/B test results. • Analyzed DAU and usage frequency differences between drive users and doc users, and visualized results at company level and industry level (Matplotlib, Seaborn). • Implemented Word2Vec and K-means algorithm to extract user preference topic by clustering monthly user data. • Established a user funnel model for WeChat Work Microdocument product, and monitored key metrics such as retention rate and conversion rate to evaluate A/B test results. • Analyzed DAU and usage frequency differences between drive users and doc users, and visualized results at company level and industry level (Matplotlib, Seaborn). • Implemented Word2Vec and K-means algorithm to extract user preference topic by clustering monthly user data.