Public Proof Lab
Reference code for regulated, agentic, healthcare AI systems.
Public repositories that make the portfolio easier to verify: RAG patterns, policy gates, execution receipts, evaluation harnesses, FHIR integration, MLOps, compliance automation, observability, and healthcare agent scaffolds.
Operating Principles
The public repos are not client code. They are reference implementations and reusable patterns that show how I think about production AI risk.
Public Repositories
Each card links directly to GitHub so reviewers can inspect the implementation surface. Evidence levels describe what kind of proof each repo provides.
healthcare-rag-platform
Production HIPAA-compliant RAG for healthcare with PHI detection, guardrails, FDA audit logging, and evaluation metrics.
Public reference implementation
model-governance-framework
AI governance for healthcare ML: bias detection, fairness metrics, FDA-ready model cards, and responsible AI controls.
Public control framework
mlops-healthcare-platform
FDA 21 CFR Part 11 compliant MLOps with IQ/OQ/PQ validation, MLflow integration, model registry, and audit trails.
Public platform pattern
compliance-automation-suite
HIPAA compliance automation with PHI validation, 45 CFR 164 references, audit logging, and security controls.
Public compliance pattern
fhir-integration-service
FHIR R4 integration service for healthcare interoperability and regulated data exchange patterns.
Public integration service
agentic-receipts
Cryptographic receipts for agent traces using hash chains, Ed25519 signatures, and JCS canonicalization.
Public verification primitive
agentic-trace-cli
CLI for canonicalizing, signing, and verifying agent execution receipts.
Public developer tool
agentic-policy-engine
Deny-by-default policy decisions with symmetric receipts at the tool boundary.
Public safety-control pattern
agentic-eval-harness
Per-scenario regression gates for bypass-resistance and deterministic agent behavior testing.
Public eval harness
agentic-evidence-viewer
Read-only viewer for agent execution bundles including trace chain, signatures, and redactions.
Public artifact viewer
agentic-member-assistant
Production-style virtual health assistant with identity-scoped retrieval, tool orchestration, and safe benefit explanations.
Public application scaffold
agentic-streaming-backend
SSE streaming backend for agentic AI with backpressure, cancellation, retry, and observability.
Public backend pattern
Reviewer Fast Path
For a 90-second review, start with healthcare RAG, governance, MLOps, and agentic receipts. For a technical deep dive, inspect the policy, eval, and streaming-backend patterns.