Data engineer building SQL and Python pipelines, data models, and reporting systems.
United States | Remote · Open to data engineering roles and project work.
I make the logic explicit, reconcile results to the source, and document the handoff. My professional background is in healthcare reporting and business intelligence. My independent work applies the same discipline to public data: clear definitions, visible checks, inspectable workflows, and limitations stated up front.
These are independent portfolio and academic projects using public or synthetic data.
I combined graph traversal and vector search to connect evidence across documents. The Streamlit app compares retrieval paths and exposes the passages behind each answer. Python, Kuzu, and LanceDB connect the document passages, relationships, and vector index.
The demo uses 11 synthetic documents, hash embeddings, and heuristic answer extraction. Its three-question benchmark illustrates the approach; it does not establish general accuracy or production performance.
Try the demo · Read the approach
I built a retrieval demo that filters synthetic hospital records by tenant, role, and clearance before ranking eligible passages with lexical cosine similarity. Change the access settings to inspect which records become eligible.
The Streamlit demo simulates access rules without authentication or database row-level security. The repository includes a separate FastAPI backend with PostgreSQL row-level security, pgvector, and tenant audit chains; that backend does not run in the demo.
Try the demo · Read the approach
- NY SPARCS Healthcare Operations BI: Python and DuckDB normalize annual discharge files across schema changes and build Parquet reporting marts. The app explores sample data, not statewide totals.
- ED Operations Analytics: Scotland's emergency-department data modeled by site and month in SQL and Python, with reconciliation checks. The forecast comparison's uncertainty interval includes zero improvement.
- Ecommerce Retention & Growth: Subscription features using data available before each prediction cutoff, with a real-data sample and a separate synthetic demo.
- AutoModeler: Financial data to Excel models with linked formulas. Stops when a balance-sheet adjustment exceeds the configured limit.
- Ticket Intel: TF-IDF and Naive Bayes ticket classification. The app uses synthetic examples; code and evaluation are in the repository.
- SignalForge: Churn-model comparison on IBM Telco data, with calibration and evaluation choices documented. Portfolio case study.
Additional work: Rural Hospital Closure Risk · Pit Wall Intelligence · A/B Test & Experimentation Analyzer
My independent data engineering practice for organizations big and small, across industries. I extract data from legacy systems and files, build quality checks and documented pipelines, and hand off code and runbooks.
A product in progress for metric definitions, named owners, written formulas, source references, and reporting links. Source references are metadata. MeasureMap does not run pipelines or calculations.
- Founder and Principal, CGC Labs (May 2026 to present). Independent practice focused on data engineering, documented pipelines, and handoff.
- Business Intelligence Analyst, critical access hospital (February 2023 to May 2026). I built daily T-SQL ELT and dimensional models, created SQL staging and transformation for an acquired clinic's migration into Paragon, added billing-data validation checks, and automated recurring reports and regulatory submissions.
- MBA and M.S. Data Science studies, Eastern University (in progress, expected 2027).
- Bachelor of Applied Science, Business Administration, Peru State College (2022).
- Earlier work in manufacturing and law enforcement taught me to find the useful signal in messy information and explain it to people making decisions.
- Professional work: T-SQL, PL/SQL, dimensional modeling, ETL and ELT, Tableau, Power Automate, Excel, and Paragon EHR data integration.
- Independent projects: Python, DuckDB, dbt, pandas, Parquet, Kuzu, LanceDB, pytest, GitHub Actions, FastAPI, Streamlit, TypeScript, Next.js, scikit-learn, XGBoost, LightGBM, and Optuna.
- Graduate coursework: Snowflake, BigQuery, Power BI, and R. I do not present these as production experience.
Portfolio · CGC Labs · MeasureMap · LinkedIn · Email