Principal AI Architect

Job Title: AI-DLC Architect

Job Location: Edina, MN (Hybrid 3 Days/Week)

The AI Architect serves as a hands-on technical design authority for enterprise AI and AI-enabled application delivery. This role is responsible for designing secure, scalable, reusable AI architectures and enabling engineering teams to move AI solutions rapidly from experimentation to production.

The role combines hands-on solution architecture with systems-level thinking, covering LLM-enabled applications, agentic systems, AI platforms, RAG, context and memory management, evaluation, observability, security, and governance.

The ideal candidate can rapidly translate business outcomes into working architecture, evaluate emerging AI technologies objectively, guide engineering teams through AI-driven development lifecycle (AI-DLC) practices, and determine where AI acceleration is appropriate versus where deterministic engineering, controls, and human validation remain essential.

Key Responsibilities

AI Solution Architecture

• Design end-to-end architectures for LLM-powered applications, agentic AI, multi-agent systems, RAG, AI-assisted workflows, and autonomous solutions.
• Define patterns that appropriately combine LLM-based reasoning with deterministic application logic for reliable and compliance-sensitive processes.
• Design reusable architectural components including agent templates, shared libraries, RAG pipelines, vector stores, hybrid search, context management, memory, and tool/API integration.
• Translate business requirements and rapid prototypes into secure, scalable, production-ready architectures.
• Provide hands-on architectural guidance during rapid prototyping and solution build sessions.

AI Platform & Engineering

• Architect reusable AI platform capabilities that accelerate delivery across engineering teams.
• Define patterns for agent registries, model management, AI Bill of Materials (AI-BOM), model provenance, versioning, and context/resource connectivity.
• Design model-routing strategies covering SLMs/LLMs, semantic caching, fallback models, latency optimization, token utilization, and cost optimization.
• Establish AI observability covering telemetry, traces, evaluation results, operational monitoring, auditability, and cost visibility.
• Support AI platform capabilities through CI/CD, automation, infrastructure, and policy-as-code.

AI-DLC & Productionization

• Define architecture aligned with an AI-native delivery loop:

Intent Generate Evaluate Refine Validate Productionize

• Establish architectural and technical exit criteria for prototype/sandbox-to-production graduation.
• Enable teams to move rapidly from idea to prototype while maintaining clear engineering standards for production readiness.
• Drive early architectural decisions to reduce ambiguity and prevent technical debt.
• Help engineering teams distinguish between what AI can accelerate and what requires traditional engineering, security, testing, and human validation.
• Create reusable reference architectures, patterns, decision trees, and accelerators that compress future delivery cycles.

AI Evaluation, Security & Governance

• Establish evaluation frameworks covering LLM-as-a-judge, golden datasets, prompt testing, regression evaluation, groundedness, hallucination testing, and human-in-the-loop validation.
• Integrate AI evaluation and validation into engineering and deployment pipelines.
• Define security and privacy architecture for PII, PHI, sensitive enterprise data, identity, authorization, and data boundaries.
• Design architectures supporting regulatory, privacy, audit, Responsible AI, and governance requirements.
• Evaluate AI frameworks, models, platforms, and protocols using objective architectural criteria rather than vendor preference.

Technical Leadership

• Serve as the technical design authority for AI solutions and engineering teams.
• Partner with Engineering, Enterprise Architecture, Security, Data, Product, and Operations teams.
• Review solution designs and provide hands-on guidance on architecture, scalability, security, reliability, and production readiness.
• Maintain reference architectures, architecture decision records, engineering standards, reusable patterns, and technical guardrails.
• Continuously evaluate emerging AI technologies and determine where they should be incorporated into the enterprise AI ecosystem.

Required Qualifications

• 10+ years of experience across software engineering, solution architecture, cloud/platform architecture, data, or AI/ML systems.
• Strong recent hands-on experience designing and delivering LLM and Agentic AI solutions.
• Experience with agent orchestration, RAG, vector databases, hybrid search, context engineering, memory, tool calling, and API integration.
• Strong understanding of application architecture, APIs, microservices, cloud-native engineering, data architecture, CI/CD, security, observability, and DevOps.
• Experience architecting AI systems on Azure, AWS, Google Cloud Platform, or equivalent enterprise cloud platforms.
• Practical understanding of LLM/AI evaluation, including LLM-as-a-judge, regression testing, prompt evaluation, grounding, and probabilistic system validation.
• Experience designing solutions that combine LLMs/agents with deterministic software components.
• Experience taking AI solutions from prototype/PoC through production hardening and deployment.
• Strong ability to create and communicate architecture blueprints, reference architectures, design patterns, technical standards, and Architecture Decision Records (ADRs).
• Ability to work directly with business stakeholders and engineering teams to rapidly turn ambiguous problems into executable solution architectures.

Preferred Experience

• Hands-on experience with modern AI development ecosystems such as Azure OpenAI, OpenAI, Anthropic Claude, Microsoft Copilot ecosystem, Semantic Kernel, LangChain/LangGraph, or equivalent technologies.
• Experience with multi-agent architectures, MCP/tool connectivity, AI gateways, model routing, and agent registries.
• Experience developing reusable AI platform components, harnesses, templates, SDKs, and engineering accelerators.
• Experience establishing AI governance, model provenance, AI-BOM, auditability, FinOps, and Responsible AI controls.
• Experience working with regulated or highly sensitive data environments such as healthcare, financial services, or life sciences.

What We Are Looking For

Someone who can:

• Architect and build, not just produce architecture diagrams.
• Move from an ambiguous business problem to a working solution architecture and prototype quickly.
• Think beyond individual AI use cases and create reusable enterprise patterns.
• Apply AI across the entire development lifecycle, not just add an LLM to an application.
• Balance speed and experimentation with security, scalability, reliability, cost, and governance.
• Understand the difference between a compelling AI prototype and a production-ready enterprise AI solution.
• Provide enough architecture upfront to accelerate teams without recreating a heavyweight Waterfall design process.

What This Role Is NOT

We are not looking for someone whose experience is primarily:

• Traditional Solution Architecture with limited hands-on AI experience.
• AI/ML engineering focused only on model development.
• Prompt engineering or RAG implementation alone.
• Creating PoCs without experience taking them to production.
• Architecture governance without hands-on engineering understanding.
• Using AI coding assistants while continuing to operate entirely through a traditional SDLC.