I build data and AI projects focused on public programs, policy, operations, and decision support.
My background in public-service program administration and regulatory work informs the questions I choose to analyze: how programs perform, where access gaps exist, how policy changes can be evaluated, how limited resources can be allocated, and how data systems can make complex information easier to use.
My portfolio combines SQL, Python, statistical analysis, machine learning, causal inference, forecasting, geospatial analytics, optimization, and applied AI.
Service Operations SQL Analytics
Relational SQL analysis of simulated public-service operations covering referrals, cases, appointments, service delivery, conversion funnels, SLAs, cohorts, repeat utilization, and management KPIs.
Public Program Compliance Monitoring & Risk Analytics
SQL and Python compliance-monitoring system for a simulated regulated program, combining provider scorecards, corrective-action monitoring, exception detection, temporal risk modeling, and review-capacity analysis.
Medicaid Unwinding & Healthcare Access
State-level causal policy evaluation using CMS and Census data, continuous-exposure difference-in-differences, event-study diagnostics, clustered inference, placebo testing, and robustness analysis.
SIPP Economic Stability & Poverty Transitions
Longitudinal Census analysis measuring poverty entry and exit using SIPP panel data, replicate-weight inference, household-linked observations, and prospective modeling.
NHIS Healthcare Access Analysis
National analysis of cost-related delayed medical care using complex-survey inference, multiple imputation, predictive modeling, calibration, and subgroup diagnostics.
NYC 311 Service Demand Forecasting
Time-series forecasting of more than 16 million NYC 311 requests using rolling-origin validation, statistical and machine-learning benchmarks, uncertainty analysis, and workload planning.
Refugee Resettlement Capacity & Resource Planning
Forecasting and optimization platform for refugee-arrival demand, staffing capacity, service coverage, constrained resource allocation, and scenario planning.
Texas Refugee Service Access Atlas
Geospatial and transit-network analysis of access to refugee-relevant services across Dallas and Tarrant Counties using Census data, GTFS schedules, routing, service-gap analysis, and facility-location optimization.
Texas Refugee Medical Assistance — Longitudinal Geospatial Analysis
Six-year synthetic longitudinal analysis of program enrollment, geographic expansion, demographic composition, persistence, and concentration across Texas ZIP codes.
Texas Afghan Support Program — Geospatial Analytics
Interactive synthetic program analytics examining participation, financial assistance, engagement, concentration, and monthly change across five Texas metropolitan areas.
Refugee Policy Navigator
Evidence-grounded RAG system for U.S. refugee-policy research using hybrid retrieval, embeddings, citations, temporal source handling, abstention logic, and automated evaluation.
U.S. Immigration Policy Intelligence
Federal regulatory intelligence platform using PostgreSQL, SQL, NLP, semantic embeddings, clustering, automated Federal Register ingestion, and FastAPI search.
Languages & Data: Python · SQL · PostgreSQL · SQLite · pandas · NumPy
Statistics & Modeling: complex surveys · multiple imputation · regression · machine learning · causal inference · forecasting · longitudinal analysis
Decision Science: optimization · scenario analysis · capacity planning · operational KPIs
Geospatial: GeoJSON · Leaflet · Folium · GTFS · transit-network analysis
Applied AI: RAG · NLP · semantic search · embeddings · hybrid retrieval · evaluation
Engineering: FastAPI · Docker · GitHub Actions · automated testing · reproducible pipelines
The projects in this portfolio are designed as reproducible analytical systems rather than isolated notebooks. Repositories include source documentation, methodological assumptions, validation, tests, and—where useful—deployed interactive applications.