Looking for a job

Zoey Zeng

@zoeyzeng

Data scientist at Wells Fargo who evaluated a GenAI chatbot and improved fallback routing accuracy by 30%.

At Wells Fargo, I analyzed more than 300,000 chatbot fallback utterances and identified five core user-intent segments. I also developed a structured evaluation framework for a GenAI conversational product, improving fallback routing accuracy by 30% and reducing guardrail misses by 20%.

At VortexNet, I analyzed the acquisition-to-activation funnel and user behavior to inform an onboarding redesign that raised conversion by 12%. I also designed A/B tests and built Power BI dashboards for product and growth teams.

At Huatai United Securities, I developed and evaluated classification models for auto-loan risk, using XGBoost, Optuna, and SHAP. At Tencent, I analyzed product experiments and user behavior; feature experimentation and iteration improved student task completion by 18% and teacher efficiency by 35%.

Experience

Work history, roles, and key accomplishments

Quantitative Analyst

Wells Fargo

May 2026 - Aug 2026 (3 months)

Analyzed 300K+ user chatbot fallback utterances using Python-based EDA, clustering, intent analysis, and risk-pattern detection to identify five core user intent segments. Built a data-driven evaluation framework for a GenAI conversational product, improving fallback routing accuracy by 30% and reducing guardrail miss rate by 20%.

Python EDA Clustering Intent Analysis Risk Detection GenAI RAG Product Analytics Experimentation SQL

Data Scientist, Product Analytics & Growth

VortexNet

Apr 2025 - Aug 2025 (4 months)

Analyzed the acquisition-to-activation funnel across 20K+ users using SQL and Python, tracing the largest drop-off to onboarding completion and informing a redesign that raised onboarding conversion by 12%. Designed and analyzed A/B tests across onboarding and core features, identifying winning variants that increased feature adoption by 14%.

SQL Python KMeans Clustering A B Testing Experimental Design Funnel Analysis Cohort Analysis Power BI Product Analytics

Data Scientist

Huatai United Securities

Apr 2024 - Aug 2024 (4 months)

Developed classification models on 37K+ auto-loan applications, applying feature engineering and class-weighted Logistic Regression, Random Forest, and XGBoost to identify high-risk segments under a ~4% default rate. Optimized XGBoost with Optuna and evaluated model performance using PR-AUC and threshold analysis.

Python Logistic Regression Random Forest XGBoost Optuna PR AUC SHAP Feature Engineering Risk Segmentation

Data Scientist, Education Product Growth & Experimentation Team

Tencent

May 2023 - Aug 2023 (3 months)

Designed and analyzed A/B experiments across 3,000+ users in 8 schools, translating insights from 200+ user interviews, app reviews, and service logs into hypotheses for engagement and product feature optimization. Improved student task completion by 18% and teacher efficiency by 35% through data-driven feature experimentation and iteration.

A B Testing Experimental Design User Behavior Analysis Funnel Analysis Product Analytics Python SQL