Agentic AI Architect

About the role

Owns the end-to-end AI agent architecture for Client's Agent Factory. Defines reusable patterns, orchestration standards, model selection, tool integration, security, and technical governance, ensuring modularity, scalability, and reuse across enterprise functions.

Responsibilities

• Design multi-agent reference architectures using Gemini Enterprise, ADK, Agent Engine, A2A, and ClientP.
• Define agent blueprints, composable design standards, and the Client Agent Library structure.
• Establish orchestration, memory, RAG, and tool-invocation strategy.
• Guide model selection (Gemini Pro / Flash / Flash-Lite) balancing cost, latency, and quality.
• Own security architecture (IAM, VPC-SC, Model Armor) and observability standards.
• Review and approve all agent-level technical designs and integration approaches.

Requirements

Mandatory (Must-Have) Skills

• Deep expertise in Gemini Enterprise, Vertex AI, ADK, Agent Engine, A2A, ClientP, Model Garden.
• Strong RAG, vector DB, embeddings, and knowledge-graph design.
• Secure enterprise-scale AI platform architecture (IAM, VPC-SC, DLP, guardrails).
• Microservices, containers (GKE/Cloud Run), API design.

Preferred (Good-to-Have) Skills

• Multi-agent systems and agentic orchestration patterns.
• Cost/latency optimization and model tiering.
• Prior GSI / large enterprise AI transformation experience.

Qualifications

• 12+ years in IT with 3+ years in AI/GenAI/agentic architecture.