Claire Song
Leveraging statistical modeling, ML, and rigorous experimentation to drive growth and product decisions
Profile
Data Scientist with 5+ years of experience applying statistical modeling and machine learning to solve critical business problems. Expert in Python, SQL, predictive modeling (XGBoost/LightGBM), user lifetime value prediction, and causal inference.
Career Timeline
- Engineered customer churn prediction model (LightGBM + feature engineering) with 0.892 AUC, lifting retention by 22% and recovering $500k+ annually
- Built automated A/B experimentation platform with CUPED variance reduction, reducing experiment turnaround from 5 days to 2 hours
- Established dynamic RFM user segmentation framework for marketing campaigns, increasing CTR by 35% with 4.8x ROI
- Created 20+ automated Tableau/PowerBI executive dashboards serving 200+ daily decision makers
- Conducted multi-touch attribution analysis to reallocate marketing budgets, reducing customer acquisition cost (CAC) by 19%
Projects
Open-source Python package implementing PSM, DID, and CUPED variance reduction for empirical data analysis.
