Professional Systems & Experience

Production AI systems in regulated environments, described at the right public-safe level.

Sanitized architecture summaries across insurance claims, healthcare quality, RAG/OCR, FHIR/HL7, predictive analytics, FWA, and MLOps. No employer-confidential data or private implementation details are exposed.

Sanitized professional systems

Enterprise AI systems I can discuss architecturally without exposing private data.

These cards connect resume outcomes to reusable system patterns: ingestion, validation, exception routing, observability, human review, deployment, and responsible AI boundaries.

Claims automation

Claims Document Automation / Autonomous Adjudication Support

LangGraph-style orchestration for document understanding, structured extraction, validation gates, exception routing, and downstream claims-system handoffs.

90% review-time reduction patternClaims review acceleration and safer automation boundaries without exposing employer data.
document intakeOCR/extractionagent graphvalidation gatesexception routingclaims handoff
LangGraph patternsAzureADLSCosmos DBRedisAKSSynapseobservability
Healthcare quality

HEDIS Evidence Extraction Pipeline

Quality-measure evidence workflow over HL7 MDM/ORU-style documents, base64 PDFs, OCR, rule validation, human review, and closure tracking.

20% automated closure liftAutomated closure improvement while preserving reviewer accountability.
HL7 MDM/ORUbase64 PDFOCRmeasure rulesevidence extractionhuman review
PyMuPDFTesseractFHIR/HL7rules enginereview queuesPower BI
Private document AI

On-Prem RAG/OCR Compliance Review Microservice

Private inference workflow for regulated document review using local SLMs, retrieval, OCR, containerized services, and analyst-facing review surfaces.

7K backlog clearance patternReview-time reduction and backlog clearance without moving private documents into public demos.
local document storeOCRchunkingretrievalGemma/Ollama/vLLMreview UI
PodmanStreamlitOllamavLLMRAGSLMsprivate inference
Predictive healthcare

Readmission / LOS / ED Utilization Predictive Models

Healthcare predictive modeling across claims, ADT events, demographics, diagnoses, medications, SDoH, risk scores, and operational features.

Utilization prioritizationCare-management prioritization, utilization insight, and governed model delivery patterns.
claims + ADTfeature storeXGBoost/stat modelsrisk outputsPower BIoperational handoff
PythonSQLXGBoostscikit-learnMLflowPower BICI/CD
FWA analytics

FWA Transportation Anomaly Detection

Fraud, waste, and abuse analytics over multi-leg trips, utilization anomalies, geospatial validation, and proactive review workflows.

18% waste reductionWaste reduction through targeted, explainable review signals.
trip eventsmulti-leg featuresESRI validationanomaly scoringreview queuewaste-reduction reporting
PythonSQLESRIanomaly featuresPower BIworkflow analytics
Healthcare analytics product

Healthcare Analytics Platform / Health Index

0-to-1 healthcare analytics platform work using composite risk scoring, care gaps, SDoH features, quality analytics, and embedded delivery.

$500K new revenue / $3M P4P impactNew revenue and client impact through operational healthcare analytics products.
member datarisk scoringcare gapshealth indexembedded dashboardcare-manager workflow
PythonSQLPower BI EmbeddedAzurerisk modelsanalytics product

Impact

Operational outcomes that inform the public system demos.

Selected outcomes from healthcare, insurance, hackathon builds, local model labs, and applied AI delivery. Public repositories are synthetic, sanitized, historical, or research/demo implementations only.

7Kcase backlog cleared

healthcare operations delivery

90%review-time reduction

claims and document review acceleration

20%closure lift

automated closure improvement

18%FWA waste reduction

fraud, waste, and abuse analytics

$3Mclient P4P impact

quality and value-based care programs

50%faster deployment cycles

MLOps and delivery modernization

Timeline

Public-safe narrative across healthcare, insurance, platforms, and research foundations.

Production Applied AI / ML Engineering

Healthcare, insurance, and regulated workflow automation

Built public-safe system patterns around claims automation, document intelligence, risk scoring, data contracts, review workflows, and production deployment discipline.

MetLife

Claims automation and document AI

High-level public framing only: claims workflow acceleration, review-time reduction, structured extraction, and automation safeguards without internal data or proprietary process details.

IEHP

Healthcare payer AI modernization

Public-safe framing around healthcare analytics modernization, care-quality programs, FWA analytics, and MLOps practices for payer-oriented AI workflows.

Hexplora

Healthcare analytics platform work

Platform-oriented healthcare analytics, model delivery, and operational reporting experience expressed without confidential implementation details.

Research and foundations

Forecasting, visualization, GPU planning, and ML templates

Earlier public projects show the base layer behind current systems work: model comparison, regression, RAG prototypes, activation visualization, infrastructure planning, and reusable ML patterns.

Recognition

Healthcare AI recognition that matches the portfolio lead story.

Global HL7 AI Challenge Winner

Transformative Impact in Healthcare

MeldRx Predictive AI Hackathon Winner

Healthcare prediction and interoperability recognition

HiCounselor / BCBS-NC Diabetes Risk Prediction Challenge Winner

Predictive modeling challenge recognition