Lead AI Architect – Client Technology

About the role

EY is seeking a Lead AI Architect to own the solution architecture for a major Oceania agentic initiative that modernises, agentifies and integrates multiple products used in professional services delivery. You will work closely with product, engineering, data, security and regional teams to ensure pragmatic, scalable and secure solutions aligned with global strategy.

Key responsibilities

• Lead the end‑to‑end architecture of the agentic solution across business process, AI, application, data, integration, security and operations.
• Define, document and maintain architecture decisions, patterns, risks and trade‑offs throughout the initiative.
• Align the solution with global architecture standards, Tax technology strategy and EY security requirements.
• Identify platform and capability gaps in Azure, integration services, data access, knowledge retrieval and drive them into the appropriate backlog.
• Drive AI governance, quality practices and production support patterns, including model selection, prompt evaluation and cost management.
• Partner with engineering to ensure designs are implementation‑ready and delivered within realistic timelines and budgets.

Required profile

• Extensive experience designing enterprise‑scale solutions on Microsoft Azure.
• Strong knowledge of generative AI, agentic architecture patterns and Retrieval‑Augmented Generation.
• Proven ability to translate business needs into pragmatic, reusable architecture designs.
• Experience working in large, multi‑regional technology programmes, preferably in professional services or tax technology.

Required skills

• Microsoft Azure cloud architecture (application, integration, identity, networking, security, monitoring).
• Azure AI Foundry and Azure OpenAI model consumption, prompt engineering and model governance.
• Generative AI and agentic patterns (orchestration, tool calling, skills, workflows, memory, planning).
• Retrieval‑Augmented Generation (document ingestion, vector indexing, grounding, citation).
• AI evaluation and quality tooling (model/prompt evaluation, hallucination testing, safety testing).
• Enterprise integration (REST APIs, event‑driven integration, workflow orchestration).