Agentic AI Architecture
Designing multi-agent systems that are safe to put in front of real policyholder data and real money. My work centers on the emerging agentic protocol stack and, just as importantly, on the boundaries — where deterministic control has to hold the line and where agentic reasoning earns its place.
Model Context Protocol (MCP) · Agent-to-Agent (A2A) protocols · LangGraph orchestration · agent identity & workload attestation (SPIFFE/SPIRE) · multi-agent system design · retrieval-augmented generation · targeted fine-tuning (Amazon Bedrock) · the deterministic–agentic boundary
AI Governance & Production Assurance
The discipline that turns "the demo worked" into "we can run this in production and defend it to a regulator." This is my signature body of work: proprietary frameworks that assess whether an agentic system is genuinely ready, where the human decision rights sit, and how it will be operated once it's live.
Production-readiness assessment · human decision & control design · agent operations · policy-as-code enforcement (OPA) · NAIC AI evaluation alignment · ISO/IEC 42001–oriented governance · PII risk classification
Technology Strategy & CTO Advisory
Advising technology leadership on the decisions that set the direction — target-state architecture, modernization, and where to place AI bets. I speak the carrier CTO/CIO's language because I've lived in it: strategy that survives contact with delivery, regulation, and budget.
Technology strategy & roadmaps · target-state architecture · build / buy / partner decisions · AI adoption strategy · modernization planning · operating-model & delivery design
Enterprise & Integration Architecture
The connective tissue. Carriers and TPAs live and die on integration, and I've spent years designing the API, messaging, and communications layers that make multi-carrier operations work — increasingly turning those same assets into revenue rather than pure cost.
Enterprise architecture · API & integration architecture · event-driven architecture · MuleSoft · Azure API Management · Service Bus · API productization & monetization · Smart Communications
Cloud Architecture & FinOps
Designing cloud platforms that are sound and affordable. Agentic and data workloads have real unit economics, and I bring both the architecture and the cost discipline — landing zones and well-architected design on one side, FinOps and workload economics on the other.
Azure · AWS · cloud landing zones & well-architected design · FinOps / cloud cost governance · AI & data workload unit economics · platform reliability
Data Platform Architecture
Agents and analytics are only as trustworthy as the data underneath them. I architect governed lakehouse platforms with the lineage and catalog controls that AI workloads and auditors both require.
Databricks medallion lakehouse · Unity Catalog & data lineage · Genie / natural-language data access · data governance · analytics enablement
Life & Annuity Domain Context
Working inside a multi-carrier L&A TPA environment (15+ carriers, 4M+ policies) gives me the operational and regulatory context most AI architects don't have. I know enough of how carrier operations and compliance actually work to design systems that fit them — and to translate between AI possibility and business reality.
L&A TPA operations · carrier & regulatory context · ACORD / NAIC familiarity