Looking for a job

Star User

@starpowerchen

Message

Data scientist and health researcher with US training.

China

Message

At Merck, I migrated a legacy Excel and R Monte Carlo forecasting model to a Python pipeline, cutting simulation runtime from overnight to under 20 minutes. I also developed oncology targeting and forecasting models that informed promotional investment and patient opportunity analysis.

Earlier at Merck, I developed deep-learning models for pediatric pneumonia imaging, including a classifier that achieved an AUC of 0.977. My work spans pharmaceutical data science, clinical machine learning, and public health research, with publications across these fields.

Experience

Work history, roles, and key accomplishments

Data Scientist, Oncology Marketing Insights

Merck

Oct 2022 - Feb 2026 (3 years 4 months)

Led migration of legacy Excel/R Monte Carlo simulation to Python, reducing runtime from overnight to under 20 minutes. Implemented marketing mix models, customer targeting propensity models, and RAG agents for biomarker testing insights.

Python R Monte Carlo Simulations Marketing Mix Modeling Random Forest SMOTE SARIMA XGBoost Positive–Unlabeled Learning RAG LLM Text To SQL Power BI GenAI Data Engineering

Scientist, Data Science and Drug Discovery

Agios Pharmaceutical

Dec 2021 - Aug 2022 (8 months)

Developed a Bayesian probabilistic framework using Naive Bayes classification on biobank data to estimate disease burden of rare genetic disorders, providing a 15% lift in identification of undiagnosed patients.

Bayesian Statistics Naïve Bayes Biobank Data Genomics Python

Postdoc/Associate Scientist, Real-World Evidence

Merck

Oct 2018 - Aug 2021 (2 years 10 months)

Developed deep learning models for pediatric pneumonia detection and lung segmentation, achieving high accuracy and external validation. Conducted quasi-experimental evaluations and ICD-10 coding validation.

Deep Learning CNN Transfer Learning U Net DICOM Python R Analysis Of Quasi Experimental Research Designs ICD 10 Real World Evidence

Graduate Researcher, Psychological Sciences & Statistics

University of Connecticut

Jan 2013 - Dec 2018 (5 years 11 months)

Conducted research on HIV risk determinants using Bayesian multilevel modeling, GEE, and longitudinal analyses. Published multiple peer-reviewed articles in behavioral medicine and substance use journals.

Bayesian Hierarchical Modeling MCMC GEE Logistic Regression Principal Component Analysis Mixed Effects Models Systematic Literature Review R SAS

Research Assistant, Pediatric Prevention Research Center

Wayne State University School of Medicine

Aug 2011 - May 2012 (9 months)

Performed statistical analyses on HIV-related risk behaviors among female entertainment workers, applying multiple imputation and hierarchical modeling. Conducted psychometric assessments and qualitative coding.