Tuyển dụng Middle Python AI Cloud Engineer Azure , PostgreSQL tại Hà Nội

ROLE OVERVIEW 

• We are looking for a proactive, skilled, and detail-oriented Python AI Cloud Engineer – Azure (Middle) with 3 to 5 years of commercial software engineering experience to expand our Platform Engineering team. In this role, you will design, build, and operate Python serverless workflows on Azure Durable Functions and production AI agents on Microsoft Foundry and LangGraph, running inside a complex, network-isolated enterprise Azure landing zone. 
• You will own long-running orchestrations, agentic workflows, and event-driven integrations end to end, connecting internal enterprise tools, Microsoft 365, and our central SAP ERP system. You will build the agent harness around our models (tools, memory, guardrails, evaluation, and observability) and work hands-on with the infrastructure underneath it all: hub-and-spoke virtual networks, private endpoints, hybrid connectivity to on-premises systems, identity, and governance. 

YOUR RESPONSIBILITIES 

• Durable Functions Development: Design and build Python workflows on Azure Functions and Durable Functions (orchestrator, activity, and entity functions) using patterns such as function chaining, fanout/fan-in, async HTTP APIs, monitors, and human interaction, with idempotent activities, retry policies, and safe versioning of long-running orchestrations. 
• Backend & API Development: Build and maintain Python services and REST APIs (FastAPI, Azure Functions HTTP triggers) exposed through Azure API Management, with typed, tested, and documented code (type hints, Pydantic, OpenAPI). 
• AI Agents & Agent Harness: Design and operate production AI agents with LangGraph and Microsoft Foundry (Foundry Agent Service, Azure OpenAI and other Foundry models), including the agent harness around them: tool / function calling against internal and SAP APIs, MCP servers, state and memory, multi-agent orchestration, human-in-the-loop approvals, and durable long-running agent execution on Durable Functions. 
• RAG, Evaluation & Responsible AI: Build retrieval pipelines on Azure AI Search (hybrid / vector search, chunking, embeddings), set up offline and online evaluations and tracing in Foundry, and apply guardrails (Azure AI Content Safety, prompt-injection defenses, data-access boundaries); manage model deployments, quotas, latency, and token cost through private, Entra ID-secured Foundry endpoints. 
• Cloud Infrastructure & Networking: Provision and operate a complex Azure landing zone with Bicep or Terraform: hub-and-spoke VNets, subnets and NSGs, Private Endpoints and Private DNS zones, VNet integration for Functions Premium / Flex Consumption plans, Azure Firewall, Application Gateway / Front Door with WAF, and hybrid connectivity (VPN Gateway / ExpressRoute) to on-premises systems, applying Azure Well-Architected Framework principles.
• Security, Identity & Governance: Secure workloads with Microsoft Entra ID, Managed Identities, Azure Key Vault, Azure RBAC, and Azure Policy; enforce least-privilege access and private-only connectivity across subscriptions and environments. 
• SAP & Enterprise Integration: Design, deploy, and monitor integration pipelines connecting internal applications to SAP S/4HANA / ECC and Microsoft 365 via OData (v2/v4), REST APIs, Azure Service Bus, Event Grid, and Logic Apps. 
• CI/CD, Quality & Observability: Build and maintain Azure DevOps or GitHub Actions pipelines for infrastructure and application deployments (environment promotion, staged / slot-based rollouts); maintain high test coverage with pytest; and implement structured logging, distributed tracing, dashboards, and alerting with Application Insights, Azure Monitor, and OpenTelemetry.

Must-have

• Python: Strong commercial experience with modern Python (3.10+), including asyncio, type hints, packaging and dependency management (uv, Poetry, or pip), and testing with pytest. 
• Azure Functions & Durable Functions: Hands-on commercial experience building and operating Azure Functions in Python, including Durable Functions orchestrations in production, with a solid understanding of their constraints (deterministic orchestrator code, replay, idempotency, task hub storage, scaling, and hosting plans).
• AI Agents & LLM Engineering: Commercial experience shipping LLM-powered applications or agents to production in Python with LangGraph (or LangChain, Microsoft Agent Framework, Semantic Kernel), including tool calling, structured outputs, RAG, and systematic evaluation, plus hands-on experience with Microsoft Foundry / Azure OpenAI model deployments.
• Azure Infrastructure & Networking: Proven experience setting up and troubleshooting non-trivial Azure environments (VNets, subnets, NSGs, Private Endpoints, Private DNS, VNet integration, routing and firewalls, hybrid connectivity) using infrastructure as code with Bicep or Terraform. 
• API Design & Integration: Proven track record in designing, documenting (OpenAPI/Swagger), consuming, and securing RESTful APIs, webhooks, and message-based integrations (Service Bus, Event Grid, Storage queues).
•  Databases: Strong practical experience with relational databases (Azure SQL Database, Azure Database for PostgreSQL), schema migrations (e.g., Alembic), query performance tuning, and Python data access (SQLAlchemy, pyodbc, or psycopg). 

Nice-to-have 

• AI-102 (Azure AI Engineer Associate) certification.
• Microsoft Azure certifications such as AZ-204 (Azure Developer Associate), AZ-104 (Azure Administrator), AZ-700 (Azure Network Engineer Associate), or AZ-400 (DevOps Engineer Expert). 
• Authentication and authorization with Microsoft Entra ID (Azure AD), app registrations, MSAL, Managed Identities, Azure RBAC, PIM, and Azure Policy. 
• Azure Landing Zones: Experience with the Cloud Adoption Framework, multi-subscription hub-andspoke or Azure Virtual WAN designs, and management-group level governance. 
• Advanced Durable Functions: Experience with the Durable Task Scheduler or alternative storage providers (MSSQL, Netherite), Durable Entities for stateful coordination, and performance tuning of high-volume orchestrations. 
• Enterprise & SAP Integration: Hands-on experience or solid working knowledge of enterprise application integration, particularly with SAP (OData v2/v4, REST endpoints, SAP Gateway, or BAPI/RFC) or similar ERPs, including via Azure Integration Services (Logic Apps SAP connector, API Management, On-premises Data Gateway). 
• Experience with Docker and containerized workloads on Azure Container Apps or AKS.
• Basic frontend experience (e.g., React) for building internal admin pages and operations dashboards.
• Agent Harness Engineering: Experience building agent runtimes or harnesses: tool sandboxing, context and memory management, retries and checkpointing, cost / token budgets, and agent observability with OpenTelemetry GenAI tracing or LangSmith.
• Experience with Model Context Protocol (MCP), Agent-to-Agent (A2A) protocols, or Microsoft 365 Copilot / Teams agent extensibility.
• Experience with private networking for AI services (Foundry and Azure AI Search behind Private Endpoints, API Management as an AI gateway with token limits and load balancing). 

CORE TECHNOLOGY STACK

• Compute & Backend: Python 3.11+, Azure Functions (Python v2 programming model), Azure Durable Functions (orchestrator, activity & entity functions), FastAPI, Azure Container Apps, Azure App Service.
•  AI & Agents: Microsoft Foundry (Az