Agentic AI Architect

Responsibilities Define VSSE/PD‑EAT target reference architecture aligned to the standard pattern (portable design, scalability, performance, production readiness). Decide where multi‑agent + supervisor orchestration is required vs. simpler patterns; align to Client routing/decision logic and propose accuracy optimizations. Establish interoperability architecture using MCP (client/server boundaries, contracts, governance model) and sequence initial integrations (e.g., app‑to‑app connectivity). Define LLM strategy (client LLM endpoint usage vs direct model endpoints), prompting standards, and evaluation approach for enterprise consistency. Define API strategy (service boundaries, versioning, API governance/registration via Apigee where required). Set non‑functional requirements : latency/throughput targets (as they get formalized), reliability, observability, cost controls, and scaling guardrails. Define DevSecOps architecture aligned to templates ( Terraform/Tekton ) and compliance gates (42Crunch/SonarQube/FOSSA). Create architecture deliverables: reference diagrams, integration patterns, NFR checklist, decision records, and runbooks for production operations. Coordinate between customer and offshore team. Hands‑on experience is required, as the candidate will also need to work directly with the customer team. #J-18808-Ljbffr Programmers.io