Enterprise AI Architect

Purpose of the role

Act as ArcelorMittal’s enterprise design authority for AI and agentic solutions. Define reusable target architectures and patterns, chair design reviews, and ensure platform, data, AI, integration, security and operability decisions converge on a coherent enterprise architecture.

Key responsibilities

• Define and maintain enterprise AI reference architectures and the approved agent/AI pattern catalogue.
• Set the technical Definition of Ready for AI use cases and assess feasibility, solution options, cost/performance and non-functional requirements.
• Chair the AI Architecture Review Board and own architecture decisions, exceptions and technical-debt registers.
• Review platform and pod designs for reuse, operability, security and alignment with ArcelorMittal standards.
• Align AI patterns with Azure AI Foundry, Copilot Studio, Microsoft Fabric, Entra ID, Azure API Management, Aura/Databricks and SAP integrations.
• Ensure evaluation, observability, fallback, identity, data and operational requirements are designed before release.
• Coach architects and pod technical leads across Europe and India and validate partner accelerators before adoption.
• Accept the architecture handover pack from Microsoft and progressively assume full design authority.

Experience

15+ years in solution or enterprise architecture, including 3+ years designing generative AI or agentic solutions on Azure.

P1 - must have

• Enterprise/solution architecture on Microsoft Azure, including Azure AI Foundry, Azure AI Search, Copilot Studio and Microsoft Fabric.
• Agentic AI architecture: decomposition, orchestration, tools, memory, evaluation and guardrail patterns.
• Integration and security architecture using Azure API Management, Entra ID, managed identities, private networking and container platforms.
• Architecture governance: review boards, Architecture Decision Records, non-functional requirements and technical-debt control.
• Ability to translate business use cases into evaluable, secure and operable solution designs.
• Executive-level communication and constructive challenge across business, architecture, risk and engineering stakeholders.

P2 - preferred

• Databricks/lakehouse and semantic-layer architecture.
• SAP integration patterns and Power Platform connectors.
• Responsible AI, EU AI Act and privacy-by-design awareness.
• Azure Solutions Architect Expert, AI-102 and/or TOGAF.

Core competencies

• Stakeholder orientation and executive influence
• Results and value orientation
• Courage to challenge and sound judgement
• Cross-cultural collaboration across Europe and India
• Learning agility and capability building
• Structured decision-making and transparent governance

Key deliverables

Enterprise AI reference architecture; AI/agent pattern catalogue; feasibility assessments; Architecture Decision Records; Architecture Review Board packs and decision register; technical dependency/debt roadmap; Microsoft architecture handover pack.

First 90 days

• Complete the Zenith 2.0, security, data and Responsible AI induction.
• Confirm Microsoft counterparts and the architecture decisions, artefacts and patterns to be transferred.
• Establish the Architecture Review Board cadence, Definition of Ready and decision registers.
• Co-chair initial reviews and assume ownership of at least one reference architecture or recurring design decision.