We are seeking a highly experienced AI Architect to lead the design and implementation of an enterprise-scale AI-powered Bordereaux (BDX) Processing Framework for the Insurance MGA and Programs business.
The architect will define and deliver a reusable, scalable framework that leverages Databricks Mosaic AI, Azure AI Services, Agentic AI, and Large Language Models (Azure OpenAI, Anthropic Claude, Google Gemini) to automate ingestion, mapping, standardization, validation, reconciliation, and onboarding of Bordereaux data from multiple MGA partners.
The role will focus on creating a metadata-driven, configurable onboarding framework capable of handling varying BDX formats while reducing manual mapping, improving data quality, and accelerating partner onboarding.
Key Responsibilities
AI & Framework Architecture
Define enterprise architecture for AI-powered Bordereaux onboarding and processing.
Design Reusable Frameworks For
- BDX ingestion
- AI-assisted mapping
- Data standardization
- Business-rule validation
- Data quality monitoring
- Reconciliation workflows
Establish architecture standards, design principles, and implementation patterns for future onboarding initiatives.
Agentic AI Solution Design
Design Intelligent AI Agents Capable Of
- Understanding new Bordereaux layouts
- Mapping source columns to target canonical models
- Identifying missing attributes and exceptions
- Recommending transformation rules
- Generating onboarding insights
Architect Human-in-the-Loop approval workflows for business users.
Define multi-agent orchestration patterns for ingestion, mapping, validation, exception handling, and approval processes.
Databricks & Azure Architecture
Lead Architecture For
- Databricks Mosaic AI
- Databricks Vector Search
- Delta Lake
- Unity Catalog
- MLflow
- Lakeflow Pipelines
- Azure OpenAI
- Azure Data Lake Storage
Define Medallion Architecture and metadata-driven processing frameworks.
Architect scalable, cloud-native solutions to support onboarding of multiple MGA partners and lines of business.
Knowledge & Retrieval Architecture
Design Retrieval Frameworks Supporting
- Data dictionaries
- Mapping repositories
- Business rules
- Historical onboarding knowledge
- Transformation patterns
Implement RAG-based solutions to enable AI-assisted onboarding and decision support.
Develop semantic knowledge layers to improve mapping accuracy and automation.
Governance & Quality
Define enterprise AI governance standards.
Establish guardrails, security controls, lineage, auditability, and compliance controls.
Define AI evaluation frameworks, model monitoring, and observability standards.
Implement Responsible AI principles and enterprise data governance.
Leadership & Stakeholder Engagement
Partner with Business SMEs, Underwriting Teams, Data Architects, and Platform Engineering teams.
Conduct architecture reviews and design workshops.
Lead POCs, framework accelerators, and reference implementations.
Mentor AI Engineers, Data Engineers, and Solution Architects.
Required Experience
12+ years of Data, Cloud, AI/ML, or Solution Architecture experience.
5+ years designing enterprise-scale Azure and Databricks solutions.
3+ years architecting Generative AI and Agentic AI solutions.
Experience building metadata-driven data platforms and onboarding frameworks.
Experience in Insurance, MGA, Programs, Bordereaux Processing, or Data Modernization initiatives preferred.
Technical Skills
AI & Agentic AI
Azure OpenAI
Anthropic Claude
Google Gemini
LangGraph
LangChain
AutoGen
CrewAI
LlamaIndex
Databricks Mosaic AI
Databricks Ecosystem
Databricks Mosaic AI
Databricks Vector Search
Delta Lake
Unity Catalog
MLflow
Lakeflow Pipelines
Databricks SQL
Azure Platform
Azure OpenAI
Azure AI Search
Azure Data Lake Storage
Azure Functions
Azure DevOps
Azure Kubernetes Service
Data Engineering
Python
PySpark
SQL
Spark
API Integrations
Metadata-driven Framework Design
Data Quality & Data Governance
Preferred Insurance Experience
Bordereaux Processing (BDX)
MGA & Programs Business
Delegated Authority Business
Policy & Claims Data Integration
Underwriting Operations
Data Modernization Programs
