Open to opportunities

Aaron Allen

@aaronallen2

Message

Senior data scientist focused on causal inference, clinical AI, and rigorous evaluation of ML and GenAI systems.

United States

Message

What I'm looking for

I’m looking for a senior data science role where I can apply causal inference and experimental rigor to healthcare or adjacent domains—especially for clinical AI and GenAI—so model evaluation translates into reliable decisions.

I’m a Data Scientist with 11 years of experience building models and analytical systems that drive real business and clinical decisions across healthcare, retail analytics, clinical AI, and housing finance. The through-line in my work is causal inference—isolating what actually caused outcomes, not just measuring what happened—starting from financial services analytics and continuing through consulting DS work.

Since joining Excella in 2020, I’ve focused deeply on healthcare: building survival analysis using Cox proportional hazards and Kaplan-Meier visualization, designing multimodal ML pipelines that combine EHR, RNA sequencing outputs, and molecular biomarker signals, and turning unstructured oncology notes into structured features via clinical NLP (spaCy and Hugging Face Transformers). I also applied DoWhy to isolate causal effects of treatment decisions in observational clinical data, aiming to deliver more reliable evidence for research teams.

More recently, I’ve brought the same experimental rigor to statistical evaluation of GenAI systems—designing LLM response accuracy, clinical relevance, and factual grounding frameworks for physician-facing workflows. Across genomics and finance, I’ve used SHAP for interpretable, auditable governance and causal methods (including CausalML) for treatment/program impact measurement, supported by reproducible MLOps with MLflow, Docker, DVC, and AWS.

Experience

Work history, roles, and key accomplishments

Senior Data Scientist

Excella

Jun 2020 - Jul 2026 (6 years 1 month)

Built genomic and clinical survival analysis models using Cox proportional hazards and Kaplan-Meier, and developed multimodal PyTorch pipelines combining EHR and RNA-seq biomarkers for treatment response prediction. Applied causal inference with DoWhy and created statistical evaluation frameworks for Tempus LLM-based clinical document tools.

Survival Analysis Cox Proportional Hazards Kaplan Meier DoWhy Causal Inference PyTorch SpaCy Hugging Face Transformers R Bioconductor AWS Sagemaker MLFlow DVC Docker SHAP XGBoost LightGBM LLM Evaluation Experimental Design

Data Scientist

Metability LLC

Nov 2016 - Apr 2020 (3 years 5 months)

Developed loyalty customer segmentation with KMeans and hierarchical clustering and built store-level demand forecasting using Prophet and statsmodels. Measured promotional lift with difference-in-differences and propensity score matching, and built healthcare risk stratification and predictive cost models using scikit-learn, XGBoost, and LightGBM with SHAP interpretability.

K Means Prophet Difference In Differences A B Testing scikit learn XGBoost LightGBM SHAP Pyspark PyTorch AWS S3 AWS RedShift SQL Pandas NumPy Scipy

Data Analyst

Navy Federal Credit Union

Jan 2015 - Sep 2016 (1 year 8 months)

Used SQL to pull and transform member datasets and produce recurring lending, account activity, and retention analyses for business and product teams. Built Tableau dashboards and used Python/Pandas to clean and analyze transaction data to identify spending and account usage patterns, supporting segmentation for outreach targeting.

SQL Tableau Python Pandas data transformation Data Analysis Data Visualization Customer Segmentation