BI Analyst · I figure out who leaves, who buys, and what to do about it.
Dual Master's (MBA + MS Data Science) at Eastern University, expected 2027. Six years in healthcare BI — clinical quality reporting, EMR migrations, and turning messy hospital data into decisions.
Open to BI / Data Analyst roles, especially in healthcare.
Self-hosted registry for hospitals and health systems to define, approve, and version-control every performance indicator. Solves the "spreadsheet chaos" problem that every quality team knows: dozens of measures across departments, regulatory submissions, and no single source of truth for what a metric means.
Role-based access (Admin / Steward / Viewer), LDAP/Active Directory integration, CSV import with validation, full audit trail with snapshots. Built to run air-gapped in a hospital network.
Stack: Next.js 15 · TypeScript · PostgreSQL 16 · Prisma · Docker · Tailwind
Production churn prediction on 7,043 telecom customers. Optuna-tuned Logistic Regression at 0.849 AUC [95% CI: 0.828, 0.869]. Quantified $1.67M annual revenue at risk with a retention ROI framework.
The interesting part: LR beat both ensemble models, and the margin was only statistically significant vs RF (p=0.016), not GB (p=0.130). Without confidence intervals, you'd tell a different story.
GitHub → • Dashboard → • Write-up →
| Project | What It Does | Result |
|---|---|---|
| E-Commerce Retention & Growth | XGBoost + SHAP on 400M+ event logs | 0.84 AUC, 484K user holdout |
| SaaS Churn Simulator | LightGBM churn model + ROI campaign simulator | 0.85 AUC, 3.2× lift |
| Ticket Intel | NLP support ticket routing — TF-IDF + Naive Bayes | 12ms p99, 90% F1 |
| AutoModeler | 3-statement financial models from a ticker symbol | Linked Excel via FastAPI |
Python TypeScript SQL scikit-learn XGBoost LightGBM PostgreSQL Next.js FastAPI Streamlit Docker Tableau Prisma
Currently exploring: MLflow, dbt, causal inference methods for retention analytics
- BI Analyst at Community Hospital (2022–2026) — Tableau dashboards, clinical quality reporting, EMR data migration
- MBA + MS Data Science at Eastern University (expected 2027)
- Previous: Manufacturing, law enforcement — learned to find patterns in noisy data and explain them to people who need a decision, not a model card