AI Application Solutions Architect

Arthur J. Gallagher · Rolling Meadows

Apply on the company website

Welcome to Gallagher - a global community of people who bring bold ideas, deep expertise, and a shared commitment to doing what’s right. We help clients navigate complexity with confidence by empowering businesses, communities, and individuals to thrive. At Gallagher, you’ll find more than a job; you’ll find a culture built on trust, driven by collaboration, and sustained by the belief that we’re better together. Whether you join us in a client-facing role or as part of our brokerage division, our benefits and HR consulting division, or our corporate team, you’ll have the opportunity to grow your career, make an impact, and be part of something bigger. Experience a workplace where you’re encouraged to be yourself, supported to succeed, and inspired to keep learning. That’s what it means to live The Gallagher Way.

The AI Solutions Architect decides the technical shape of AJ Gallagher's AI solutions before they are built, and makes sure that shape holds up once they are live. Reporting to the Director of AI Technology, you will take business problems from divisons or departments and turn them into architectures the team can deliver - choosing the right pattern (RAG, agent, workflow, or no AI at all), the right platform (pro-code on Microsoft Foundry or low-code in Copilot Studio), and the right integration, identity, and data design for a regulated, global insurance business. You will own the reference architectures and architecture decisions that Gallagher's AI solutions build on, partnering with the senior AI Engineering, who owns engineering standards and delivery, and with Gallagher's enterprise architecture, security, and data teams. This is a hands-on technical role for someone whose curiosity runs deep: you will prototype to test an idea rather than diagram it from a distance, and you will follow a fast-moving Azure AI platform closely enough to know which new capability solves a real Gallagher problem and which one isn't ready yet.

• Shape Solutions From Business Problems: Lead technical discovery with business stakeholders and the AI Business Analyst, turning ambiguous requests into solution options with clear trade-offs, feasibility, effort, and run-cost estimates - including recommending against AI when a simpler answer fits better.
• Own Reference Architectures: Define and maintain Gallagher's reference architectures for RAG, single and multi-agent systems, and Copilot extensibility on Microsoft Foundry, Azure OpenAI, Azure AI Search, Cosmos DB, and Azure compute, so every new solution starts from proven, approved patterns.
• Make, Record, and Explain Architecture Decisions: Make the platform and pattern calls — model selection, retrieval design, agent setup, datastore choice, messaging, hosting — and record them as architecture decision records (ADRs) with the reasoning, the alternatives, and the conditions under which the decision should be revisited; explain them to executives and engineers in terms each can act on.
• Design Governance and Data Residency In: Build Gallagher's compliance requirements into the architecture itself - Entra ID/RBAC and managed identity, Azure Policy, private networking, Standard (Regional) or DataZone Standard model deployments, Foundry agent setup with Gallagher-owned storage, audit logging, and human-in-the-loop checkpoints aligned to internal AI governance expectations.
• Prove Ideas With Working Prototypes: Build proofs of concept and technical spikes in Python and C#/.NET that test a design's riskiest assumptions - retrieval quality, latency, cost, platform limits - before the team commits to building it.
• Run Architecture Reviews: Review solution designs from AI developers, the AI Workflow Architect, and partner teams; catch integration, security, scalability, and cost problems early; and represent the AI program in Gallagher's enterprise architecture and security review forums.
• Design Enterprise Integrations: Architect how AI solutions connect to Gallagher's systems of record and data platforms through Azure API Management, Service Bus, Azure Functions, MCP servers, and Snowflake, within the enterprise identity model.
• Set the Pro-Code / Low-Code Boundary: Partner with the AI Workflow Architect to define when a solution belongs in Copilot Studio and Power Platform and when it needs a pro-code Foundry build, and design the integration points that let the two work as one solution.
• Evaluate Emerging AI Technology: Track new models, Foundry capabilities, agent frameworks (Microsoft Agent Framework, LangGraph, CrewAI), and agent protocols (MCP, A2A); run time-boxed evaluations against real Gallagher use cases; and turn findings into clear adopt / trial / hold recommendations for the Director of AI Technology.
• Design for Quality, Cost, and Operability: Specify how each solution will be evaluated (groundedness, accuracy, latency), observed in production, and kept within budget - capacity and token-cost models, quota planning, and model-version lifecycle - so designs are supportable after launch, not just at go-live.

• Experience: 10+ years in software engineering or solution architecture, including 3+ years owning architecture for enterprise cloud solutions and 2+ years designing LLM/GenAI systems - RAG, agents, or Copilot extensions. Systems that have reached production.
• Azure AI Platform Depth: Hands-on production experience with Microsoft Foundry, Azure OpenAI Service, and Azure AI Search, including retrieval design (chunking, embeddings, hybrid and semantic ranking) and agent design on Foundry Agent Service.
• Azure Architecture: Strong command of Azure compute (Functions, Container Apps, AKS, App Service), integration (API Management, Service Bus), data (Cosmos DB, Blob Storage, Azure SQL), and networking and security (VNets, Private Link, Key Vault, managed identity).
• Identity and Governance: Experience designing within an enterprise identity and control model - Entra ID, Azure RBAC, Azure Policy, data residency, audit logging - and defending those designs through security and risk review.
• Hands-On Engineering: Able to build working prototypes and read production code in Python and C#/.NET; comfortable in a codebase, not only at a whiteboard and design diagramming.
• Architecture Practice: Track record of producing reference architectures, ADRs, and solution designs that teams actually built from, and of leading architecture reviews across multiple concurrent initiatives.
• Evaluation Literacy: Experience defining how an AI system's quality is measured - test sets, groundedness and accuracy metrics, regression checks, rather than judging it by demo.
• Communication: Exceptional written and verbal ability to explain architecture trade-offs to senior business leaders and engineers alike.
• Education: Bachelor's degree in Computer Science, Engineering, Information Systems, or a related field, or equivalent practical experience.

Preferred Qualifications

• Regulated Industry Experience: Insurance, financial services, or healthcare, with exposure to NAIC AI model governance, model risk management, or SOX-type controls.
• Agent Frameworks and Protocols: Microsoft Agent Framework, Semantic Kernel, LangChain/LangGraph, or AutoGen, plus MCP and A2A integration patterns.
• Copilot Extensibility: M365 Copilot extensibility, Copilot Studio, and Agent 365, including how pro-code agents surface in Teams and M365 Copilot.
• Infrastructure-as-Code: Designing for Terraform- or Bicep-deployed infrastructure, with policy-as-code guardrails.
• AI Cost and Capacity Management: Provisioned throughput versus pay-as-you-go trade-offs, quota and throughput planning, and token-cost modeling across environments.
• LLMOps and Observability: Automated evaluation pipelines, red-teaming, tracing with Application Insights or OpenTelemetry, and drift and cost monitoring.
• Data Platform Integration: Snowflake or similar platforms feeding AI solutions, including Entra ID federation.
• Spec-Driven and AI-Assisted Development: Using spec-driven workflows and AI coding assistants to move from design to working software quickly.
• Certifications: AZ-305 (Azure Solutions Architect Expert), AI-103 (Azure AI Apps and Agents Developer Associate), or equivalent demonstrated architecture ownership.

Core Competencies

• Relentlessly Curious: Wants to know how things actually work - reads the docs, the release notes, and the source, and runs the experiment rather than trusting the marketing; treats a platform limit or an unexpected result as the interesting part of the job.
• Evidence Over Opinion: Backs architecture decisions with prototypes, benchmarks, and evaluation data, and changes course quickly when the evidence says to.
• Systems Thinker: Designs for the whole path - business process, agent, integrations, data, identity, and operations - and anticipates where the seams will break.
• Pragmatic About Technology: Adopts new tools only when they solve a real Gallagher problem, and will say when the simplest answer, or no AI at all, is the right one.
• Governance as a Design Input: Treats security, data residency, and auditability as architecture requirements, not reviews to get through.
• Clear Translator: Explains complex systems simply to executives and precisely to engineers; writes things down.

Technology Stack

• Microsoft Foundry / Foundry Agent Service - primary platform for pro-code agents and model deployment
• Azure AI Search - retrieval and grounding for RAG solutions
• Azure Cosmos DB / Blob Storage - agent state and memory, operational data, and document storage
• Azure Functions / Container Apps / AKS - compute for AI services and orchestration
• Azure API Management / Service Bus - secure API exposure and reliable messaging for multi-step agent workflows
• Microsoft Agent Framework / LangGraph - agent orchestration
• Microsoft Entra ID / Azure RBAC / Azure Policy / Key Vault - identity, access, and governance controls
• Copilot Studio / M365 Copilot / Microsoft Teams - low-code agents and deployment surfaces
• Snowflake - enterprise data feeding AI solutions
• Application Insights / Azure Monitor - tracing, observability, and cost signals
• Azure DevOps - backlog, pipelines, and architecture documentation

At Gallagher, we believe supporting our colleagues goes far beyond the role itself. For more information, visit our Benefits page.

• Competitive compensation
• Comprehensive benefits programs designed to support your well-being 
• Career development opportunities and ongoing learning 
• A collaborative, people-first culture with accessible leadership 
• The opportunity to do meaningful work with global reach and local impact 

At Gallagher, we are dedicated to building an inclusive and authentic workplace. If your past experience doesn’t align perfectly, we encourage you to join our Talent Community to stay connected to additional career opportunities. At times, we will consider transferable skills from previous roles.

Gallagher is an affirmative action/equal opportunity employer (Minorities/Females/Veterans/Disabled)

• Shape Solutions From Business Problems: Lead technical discovery with business stakeholders and the AI Business Analyst, turning ambiguous requests into solution options with clear trade-offs, feasibility, effort, and run-cost estimates - including recommending against AI when a simpler answer fits better.
• Own Reference Architectures: Define and maintain Gallagher's reference architectures for RAG, single and multi-agent systems, and Copilot extensibility on Microsoft Foundry, Azure OpenAI, Azure AI Search, Cosmos DB, and Azure compute, so every new solution starts from proven, approved patterns.
• Make, Record, and Explain Architecture Decisions: Make the platform and pattern calls — model selection, retrieval design, agent setup, datastore choice, messaging, hosting — and record them as architecture decision records (ADRs) with the reasoning, the alternatives, and the conditions under which the decision should be revisited; explain them to executives and engineers in terms each can act on.
• Design Governance and Data Residency In: Build Gallagher's compliance requirements into the architecture itself - Entra ID/RBAC and managed identity, Azure Policy, private networking, Standard (Regional) or DataZone Standard model deployments, Foundry agent setup with Gallagher-owned storage, audit logging, and human-in-the-loop checkpoints aligned to internal AI governance expectations.
• Prove Ideas With Working Prototypes: Build proofs of concept and technical spikes in Python and C#/.NET that test a design's riskiest assumptions - retrieval quality, latency, cost, platform limits - before the team commits to building it.
• Run Architecture Reviews: Review solution designs from AI developers, the AI Workflow Architect, and partner teams; catch integration, security, scalability, and cost problems early; and represent the AI program in Gallagher's enterprise architecture and security review forums.
• Design Enterprise Integrations: Architect how AI solutions connect to Gallagher's systems of record and data platforms through Azure API Management, Service Bus, Azure Functions, MCP servers, and Snowflake, within the enterprise identity model.
• Set the Pro-Code / Low-Code Boundary: Partner with the AI Workflow Architect to define when a solution belongs in Copilot Studio and Power Platform and when it needs a pro-code Foundry build, and design the integration points that let the two work as one solution.
• Evaluate Emerging AI Technology: Track new models, Foundry capabilities, agent frameworks (Microsoft Agent Framework, LangGraph, CrewAI), and agent protocols (MCP, A2A); run time-boxed evaluations against real Gallagher use cases; and turn findings into clear adopt / trial / hold recommendations for the Director of AI Technology.
• Design for Quality, Cost, and Operability: Specify how each solution will be evaluated (groundedness, accuracy, latency), observed in production, and kept within budget - capacity and token-cost models, quota planning, and model-version lifecycle - so designs are supportable after launch, not just at go-live.

• Experience: 10+ years in software engineering or solution architecture, including 3+ years owning architecture for enterprise cloud solutions and 2+ years designing LLM/GenAI systems - RAG, agents, or Copilot extensions. Systems that have reached production.
• Azure AI Platform Depth: Hands-on production experience with Microsoft Foundry, Azure OpenAI Service, and Azure AI Search, including retrieval design (chunking, embeddings, hybrid and semantic ranking) and agent design on Foundry Agent Service.
• Azure Architecture: Strong command of Azure compute (Functions, Container Apps, AKS, App Service), integration (API Management, Service Bus), data (Cosmos DB, Blob Storage, Azure SQL), and networking and security (VNets, Private Link, Key Vault, managed identity).
• Identity and Governance: Experience designing within an enterprise identity and control model - Entra ID, Azure RBAC, Azure Policy, data residency, audit logging - and defending those designs through security and risk review.
• Hands-On Engineering: Able to build working prototypes and read production code in Python and C#/.NET; comfortable in a codebase, not only at a whiteboard and design diagramming.
• Architecture Practice: Track record of producing reference architectures, ADRs, and solution designs that teams actually built from, and of leading architecture reviews across multiple concurrent initiatives.
• Evaluation Literacy: Experience defining how an AI system's quality is measured - test sets, groundedness and accuracy metrics, regression checks, rather than judging it by demo.
• Communication: Exceptional written and verbal ability to explain architecture trade-offs to senior business leaders and engineers alike.
• Education: Bachelor's degree in Computer Science, Engineering, Information Systems, or a related field, or equivalent practical experience.

Preferred Qualifications

• Regulated Industry Experience: Insurance, financial services, or healthcare, with exposure to NAIC AI model governance, model risk management, or SOX-type controls.
• Agent Frameworks and Protocols: Microsoft Agent Framework, Semantic Kernel, LangChain/LangGraph, or AutoGen, plus MCP and A2A integration patterns.
• Copilot Extensibility: M365 Copilot extensibility, Copilot Studio, and Agent 365, including how pro-code agents surface in Teams and M365 Copilot.
• Infrastructure-as-Code: Designing for Terraform- or Bicep-deployed infrastructure, with policy-as-code guardrails.
• AI Cost and Capacity Management: Provisioned throughput versus pay-as-you-go trade-offs, quota and throughput planning, and token-cost modeling across environments.
• LLMOps and Observability: Automated evaluation pipelines, red-teaming, tracing with Application Insights or OpenTelemetry, and drift and cost monitoring.
• Data Platform Integration: Snowflake or similar platforms feeding AI solutions, including Entra ID federation.
• Spec-Driven and AI-Assisted Development: Using spec-driven workflows and AI coding assistants to move from design to working software quickly.
• Certifications: AZ-305 (Azure Solutions Architect Expert), AI-103 (Azure AI Apps and Agents Developer Associate), or equivalent demonstrated architecture ownership.

Core Competencies

• Relentlessly Curious: Wants to know how things actually work - reads the docs, the release notes, and the source, and runs the experiment rather than trusting the marketing; treats a platform limit or an unexpected result as the interesting part of the job.
• Evidence Over Opinion: Backs architecture decisions with prototypes, benchmarks, and evaluation data, and changes course quickly when the evidence says to.
• Systems Thinker: Designs for the whole path - business process, agent, integrations, data, identity, and operations - and anticipates where the seams will break.
• Pragmatic About Technology: Adopts new tools only when they solve a real Gallagher problem, and will say when the simplest answer, or no AI at all, is the right one.
• Governance as a Design Input: Treats security, data residency, and auditability as architecture requirements, not reviews to get through.
• Clear Translator: Explains complex systems simply to executives and precisely to engineers; writes things down.

Technology Stack

• Microsoft Foundry / Foundry Agent Service - primary platform for pro-code agents and model deployment
• Azure AI Search - retrieval and grounding for RAG solutions
• Azure Cosmos DB / Blob Storage - agent state and memory, operational data, and document storage
• Azure Functions / Container Apps / AKS - compute for AI services and orchestration
• Azure API Management / Service Bus - secure API exposure and reliable messaging for multi-step agent workflows
• Microsoft Agent Framework / LangGraph - agent orchestration
• Microsoft Entra ID / Azure RBAC / Azure Policy / Key Vault - identity, access, and governance controls
• Copilot Studio / M365 Copilot / Microsoft Teams - low-code agents and deployment surfaces
• Snowflake - enterprise data feeding AI solutions
• Application Insights / Azure Monitor - tracing, observability, and cost signals
• Azure DevOps - backlog, pipelines, and architecture documentation

More jobs at Arthur J. Gallagher · Software Engineering jobs in Rolling Meadows