Microsoft AI Architect

OverviewRole: Microsoft AI ArchitectLocation: Onsite in Abu DhabiSalary: AED 28,000-30,000 per month + VISA + HealthcareOur client is looking for a Microsoft AI Architect to lead the enterprise-scale architecture of their Agentic AI platform. This role blends distributed systems, applied AI, and platform engineering to deliver secure, scalable, production-grade GenAI solutions.The ideal candidate is tool-agnostic and architecture-focused, with the ability to combine Microsoft’s AI ecosystem with best-in-class open-source frameworks.Experience requiredCore Engineering8+ years in backend, platform, or distributed systemsStrong expertise in Python and/or C#Experience building APIs, microservices, and event-driven systemsGenAI and Agent Systems3+ years building production-grade GenAI solutionsExperience with agent frameworks, RAG, tool-calling, and agent memoryData and SearchExperience with Azure AI Search or vector databasesStrong understanding of embeddings, hybrid search, and retrieval pipelinesAI Safety and ObservabilityExperience implementing guardrails, policy engines, and LLM monitoringTracing, evaluation, and telemetry for AI systemsLLMOps and InfrastructureCI/CD for AI workloadsModel experimentation and trackingKey ResponsibilitiesAgentic System ArchitectureArchitect and implement multi-agent systems using frameworks such as Semantic Kernel, Microsoft Agent Framework, LangGraph, CrewAI, or AutoGenDesign hybrid orchestration combining agentic reasoning and deterministic workflowsImplement agent-to-agent communication and Model Context Protocol for tool discovery and invocationRetrieval, Reasoning, and Knowledge SystemsBuild enterprise-grade RAG and Agentic-RAG pipelines using Azure AI Search or open-source vector platforms such as LlamaIndex, Haystack, Milvus, Weaviate, or QdrantEnable multi-hop reasoning across structured and unstructured dataIntegrate with Microsoft Fabric, Dataverse, Cosmos DB, and enterprise data lakesAI Governance, Safety, and ObservabilityDesign governance layers for prompt filtering, PII protection, content moderation, and policy enforcementImplement observability for prompts, agent actions, cost, latency, and riskUse tools such as Microsoft Purview, AI Content Safety, Guardrails AI, NeMo Guardrails, Arize Phoenix, LangSmith, MLflow, or OpenTelemetryLLMOps and Model InfrastructureOwn the full LLM lifecycle including prompt flows, routing, evaluation, versioning, and deploymentWork with Azure OpenAI and open-source models such as Llama, Mistral, and QwenBuild CI/CD pipelines using GitHub Actions, Azure DevOps, MLflow, BentoML, or similarDesign scalable inference on Kubernetes and cloud using vLLM, TGI, KubeRay, SkyPilot, or managed servicesHuman-in-the-Loop InterfacesBuild real-time AI interfaces using React and TypeScriptEnable streaming responses via WebSockets or Server-Sent EventsCreate dashboards for approval, override, and guidance of agent actionsIntegrate solutions into Microsoft Teams, web portals, developer tools, and VS Code AI Toolkit#J-18808-Ljbffr