Job Description
The AI Domain Architect is a senior enterprise architecture role responsible for designing and operationalizing AI-enabled solutions across assigned domains and delivery portfolios.
This role sits at the point where enterprise AI strategy becomes working software. The AI Domain Architect translates reference architectures, governance requirements, and platform capabilities into production-ready solution designs, and then stays in the work alongside engineering to see those designs through to delivery.
Operating at the enterprise level, the AI Domain Architect influences architectural direction across multiple initiatives while partnering closely with engineering, product, governance, and platform stakeholders to ensure AI systems are secure, observable, cost-effective, and aligned with AGS standards. This is not a purely advisory role. The AI Domain Architect works directly with the AI Product, Engineering, and delivery teams to ensure designs move successfully from concept to production.
Responsibilities
Solution Architecture for AI-Enabled Systems
- Design AI-enabled architectures across assigned domains, aligned with enterprise standards, platform capabilities, and AGS reference patterns
- Translate enterprise reference architectures into domain-level implementation blueprints that engineering teams can execute against
- Guide design decisions for agentic systems, orchestration workflows, retrieval and grounding patterns (RAG), model integration, and tool use
- Architect secure integration patterns covering identity, permissions, auditability, and service-to-service authentication for AI workloads
- Partner with engineering leads to ensure designs are scalable, maintainable, and realistic for the team’s context
Production Readiness and Operationalization
- Embed observability, telemetry, evaluation, and monitoring requirements into solution designs from day one
- Build in lifecycle management, model and prompt versioning, cost monitoring, and safe deployment practices
- Define evaluation approaches including baseline test sets, regression coverage, quality thresholds, and human-in-the-loop checkpoints where appropriate
- Integrate logging, safety signals, and performance metrics in partnership with AI Product, Engineering, and Delivery teams
- Support production readiness reviews and architectural risk assessments before go-live
Governance, Risk, and Responsible AI
- Translate governance and risk requirements into practical architectural controls that don’t slow delivery unnecessarily
- Ensure required documentation, evidence, and compliance checkpoints are built into delivery workflows rather than bolted on at the end
- Guide teams through architecture reviews and governance intake, including the judgment calls that sit between them
- Embed approved guardrails, content safety controls, and platform policies into solution designs
- Proactively surface architectural and AI-specific risks (data leakage, prompt injection, model misuse, cost exposure) and propose mitigations
Patterns, Reuse, and Platform Feedback
- Promote reuse of approved templates, patterns, and reference implementations across the domain
- Identify duplication and drift within the domain and recommend consolidation
- Provide structured feedback to enterprise architecture and platform teams so reusable assets keep getting better
- Contribute to evolving standards based on implementation learnings and post-production insights
Delivery Partnership and Technical Leadership
- Operate as a trusted technical partner to engineering, product, governance, and domain leadership
- Participate in technical design workshops, delivery planning, and architectural due diligence for AI-enabled integrations and third-party solutions
- Support roadmap planning by bringing architectural feasibility, scalability, and total-cost considerations into the conversation early
How You Work
- Partner, not gatekeeper. You earn the right to be heard by being in the work, not by standing outside it.
- Accelerator, with judgment. You speed teams up when you can, and slow them down when you should. You know the difference, and you can explain it.
- Hands-on when it helps. You stay current enough to prototype, pair, and debug alongside engineers, even though shipping the final code is the team’s job, not yours.
- Enterprise-minded, delivery-oriented. You hold the bigger picture without losing sight of what’s actually shipping.