IT Architect

Job Title: IT Architect- AI FLSA Status: Exempt

Job Overview

IT Architect – AI plays a key role in leading AI enablement and Azure solution design across engineering and the wider business at American Trailer World (ATW). The role sets the standard for how employees work alongside AI, delivers AI agents that streamline the software development lifecycle, and supports business teams in building their own AI tools within approved guardrails.

The position combines hands-on technical delivery with coaching and governance. All solutions are expected to follow the ATW Azure reference architecture and security standards.

.

Job Responsibilities

• Define how teams use AI in daily work and publish playbooks, agent patterns, and reusable templates.
• Design and deliver AI agents for ticket refinement, code generation, code review, test generation, CI/CD triage, and documentation, integrated through APIs and MCP servers.
• Track cycle time, review turnaround, defect escape rate, and hours saved, and refine solutions based on results.
• Partner with business teams to identify repetitive work suitable for automation, and mentor employees as they build AI agents and automations within approved guardrails.
• Set Azure standards for model selection, orchestration, retrieval, evaluation, cost, and observability for all internal and vendor-built applications.
• Guide Azure solution design, review code and pull requests, and hold vendor deliverables to the same standards.
• Establish guardrails and evaluation practices, and work with Security and Compliance on data handling, access controls, and human review.
• Lead workshops, office hours, and pairing sessions to build AI skills across the organization.
• Write and review production-quality code and reusable components, and troubleshoot complex application, integration, identity, performance, and deployment issues.
• Within the first 90 days, map the areas where AI can remove the most friction, deliver one production agent with a measurable baseline, and publish version 1 of the AI coworking playbook.
• Within the first 6 months, have several SDLC agents in daily use and at least three business teams using AI tools built with this role's support.
• Within the first year, establish AI-assisted work as the default across engineering, with reported gains in delivery speed and quality.
• Other duties, as assigned.

Education – Experience Required

• Bachelor's degree in Computer Science, Software Engineering, Information Systems, or a related field, or equivalent professional experience.
• 7+ years of software engineering experience, including at least 2 years as an architect or technical lead.
• Proven record of delivering AI agents that automate real work, with measurable results.
• Hands-on experience building MCP servers that connect agents to internal systems, including authentication and permissions.
• Strong knowledge of LLM patterns: prompt design, tool calling, retrieval-augmented generation (RAG), agent orchestration, and evaluation.
• Daily use of agentic coding tools such as Claude Code or GitHub Copilot.
• Production experience with Azure, including Azure AI Foundry or Azure OpenAI, Functions, Container Apps, Entra ID, Key Vault, Azure DevOps, and Application Insights, using .NET (C#) or Python.
• Experience leading code reviews and setting coding standards for development teams, including vendors.
• Experience helping non-technical employees adopt AI and build their own tools.
• Strong understanding of AI security, including data privacy, access control, prompt injection risks, and human oversight.
• Clear communication skills, with the ability to explain technical architecture to engineers and automation ideas to business teams such as Finance.
• Experience with Claude Cowork, Claude API, Copilot Studio, or Power Automate is preferred.
• Experience running an AI center of excellence is preferred.
• Microsoft certification such as AZ-305 (Azure Solutions Architect Expert) or AI-102 (Azure AI Engineer Associate) is preferred.
• Knowledge of Domain-Driven Design and infrastructure as code using Bicep is preferred.