← Home All Case Files ▦
SCROLL 000%
DOWNTIME: ZERO
01Home 02Case Files 03About 04Skills 05Contact ↓Resume Hire Me ↗
// CASE FILE 06 — CODENAME: SHIELD WALL — ENTERPRISE · FORTUNE 500

FMGLOBAL, ZERO-DOWNTIME GENAI.

Led the GenAI platform modernisation for FMGlobal — migrating critical insurance workloads to secure Azure AI infrastructure with zero downtime, then arming the platform with guardrailed LLM endpoints, enterprise RAG, and live MCP tool-use. Insurance is the business of holding the line. So was this migration.

Inspired By
Captain America
Role
GenAI Lead
Client
Fortune 500 Insurer
Downtime
Zero
Migrate★Guardrail★Ground★Govern★Hold the Line★ Migrate★Guardrail★Ground★Govern★Hold the Line★
FILE 01 — THE MISSION

MODERNISE THE PLANE WHILE IT'S FLYING

Insurance workloads don't get maintenance windows. The mandate: move FMGlobal's critical insurance workloads onto secure Azure AI infrastructure — and modernise the GenAI platform around them — without a minute of downtime.

Delivered as lead: the zero-downtime migration itself, multimodal LLM endpoints on Azure AI Foundry with inference auto-scaling and Azure AI Content Safety guardrails, enterprise RAG grounded in Azure AI Search hybrid indexing, live MCP tool-use against Azure Boards — and the CI/CD governance an enterprise actually audits.

"Fortune 500 trust is won in the cutover."

  • ENGAGEMENT FILE
  • Zero-downtime — staged, validated migration
  • Guardrails — Azure AI Content Safety
  • RAG — vector + semantic hybrid indexing
  • MCP — Azure Boards, live work items
  • Governance — 20+ microservices, JFrog gates
  • Note — client internals withheld
FILE 02 — WHAT SHIPPED

FIVE FRONTS, ONE PLATFORM

01

Zero-Downtime Migration

Critical insurance workloads moved to secure Azure AI infrastructure with no service interruption — staged cutovers, rollback paths, and validation at every step.

02

Guardrailed LLM Endpoints

Multimodal LLMs deployed on Azure AI Foundry — configured endpoints, inference auto-scaling, and Azure AI Content Safety standing in front of production traffic.

03

Enterprise RAG

Document ingestion, chunking, and vector + semantic hybrid indexing in Azure AI Search — measurably better grounded responses for Fortune 500 workloads.

04

MCP for Azure Boards

Model Context Protocol integration giving AI assistants live access to tickets, sprint data, and work items — real enterprise agentic tool-use, not a demo.

05

CI/CD Governance

Standardised pipeline governance across 20+ microservices, with JFrog Artifactory and Docker image promotion gates for container supply-chain auditability.

FILE 03 — THE PLATFORM

SECURE BY ARCHITECTURE

Enterprise Users

Insurance Workflows

AI Assistants

Agentic Tool-Use

▼ ▼ ▼
Content Safety Guardrails

Every Request Screened

▼ ▼ ▼
Azure AI Foundry

Multimodal LLM Endpoints · Auto-Scaling

Azure AI Search

Hybrid Vector + Semantic Index

MCP Server

Azure Boards · Live Work Items

▼ ▼ ▼
CI/CD Spine

20+ Microservices · JFrog Promotion Gates

Enterprise engagement — architecture shown, client internals withheld
0
Minutes of Downtime
20+
Microservices Governed
100%
Production Endpoints Guardrailed
END OF FILE

MISSION LOGGED. RETURN TO BASE.