Architect (Multi-skill) — AI Transformation Core Team CPG Remote

• Location: Cincinnati OH
• Remote
• Domain: Consumer Goods Experience
• Grade: P7 / P8.

Own the end-to-end technical architecture for the Foundational AI Platform agentic orchestration MCP server design and MCP gateway integration observability/SRE ITSM/AIOps and AI solutions targeting applications hosted on Azure and GCP. Drive joint architecture decisions with the Knowledge Fabric team the clients AI platform teams and the partner ecosystem.

Key responsibilities

Foundational AI Platform (own)

• Architect the agent orchestration layer multi-agent topologies planning/reasoning patterns agent-to-agent communication human-in-the-loop gates fallback and recovery.
• Define the Cockpit / control plane agent registry version control deployment runtime configuration policy enforcement traceability dashboards.
• Define the Guardrails & Responsible AI architecture PII redaction content safety jailbreak resistance output validation policy-as-code model risk management.
• Design the agent observability and evaluation architecture tracing token/cost telemetry golden-set regression suites drift detection A/B harnesses.
• Architect Agent & Copilot Starter Kits reusable templates spanning AIOps (triage incident commander RCA resolution SOP postmortem change-impact) engineering productivity and domain-specific copilots.

MCP server MCP gateway (own key area)

• MCP server reference design (Resources Tools Prompts Sampling) and an implementation playbook for service-line teams.
• MCP gateway: discovery registry routing transport bridging (stdio / Streamable / SSE) multi-tenancy.
• AuthN/AuthZ via OAuth 2.1 OIDC On behalf of execution

Knowledge Fabric collaboration (consume not build)

• Define retrieval schema and SLA contracts with the KF team. Integrate GraphRAG outputs into agents; feedback quality signals.

AI solutions on Azure GCP application estates

• Architect agentic and RAG solutions using Azure (Azure OpenAI AI Foundry AI Search Cosmos DB AKS) and GCP (Vertex AI Gemini BigQuery GKE) services native to the apps being served.
• Cross-cloud connectivity hybrid retrieval model/inference routing secrets and identity propagation across estates.

Engineering excellence and leadership

• Architecture reviews design standards mentoring pods across business units; co-own roadmap with the joint PMO; drive AI-native SDLC adoption (spec-driven frameworks AI coding assistants).

Must-have

• 10 years in architecture; 3 years experience in GenAI/agentic systems (2 years live production experience).
• Hands-on MCP server implementation and gateway integration.
• Agent frameworks (LangGraph AutoGen Semantic Kernel or equivalent) RAG/GraphRAG evals guardrails.
• Azure and GCP AI/data stacks; Kubernetes; IaC.
• Strong API security (OAuth 2.1/OIDC/mTLS) and observability fundamentals.

Good-to-have

• AIOps / ITSM integration (ServiceNow observability platforms).
• Spec-driven dev (BMAD Spec Kit OpenSpec) and AI coding assistants.
• Multi-vendor delivery in regulated enterprises.