Azure AI Engineer

Azure AI Engineer

The Azure AI Engineer is responsible for the end-to-end implementation and deployment of enterprise AI solutions on the Azure Stack. You will take ownership of building, integrating, and operationalizing AI workloads using Azure AI Foundry, Azure Databricks, Azure Data Lake, and the broader Microsoft AI ecosystem — including the design and enforcement of guardrails for responsible, secure, and compliant AI.

This is a hands-on engineering role focused on delivery: turning architectural designs into production-ready AI services, owning the deployment lifecycle, and ensuring solutions are robust, observable, and aligned with enterprise security and governance standards

Responsibilities

AI Solution Implementation on Azure

• Solution Build: Implement AI solutions on Azure AI Foundry — including agent design, model selection, prompt flows, evaluation pipelines, and deployment of fine-tuned and base models.
• Generative AI & RAG: Build retrieval-augmented generation (RAG) pipelines using Azure AI Search, Azure OpenAI, and vector stores; integrate with enterprise data sources via Azure Data Lake and Databricks.
• Model Deployment: Deploy models and AI endpoints to Azure Machine Learning, Azure AI Foundry, AKS, and Azure Container Apps; manage endpoint scaling, versioning, and traffic routing.
• Integration: Integrate AI services with downstream applications via REST APIs, Azure API Management, Logic Apps, and Function Apps.

Data Engineering on Azure Databricks & Data Lake

• Databricks Workloads: Build and operationalize data pipelines, feature engineering jobs, and model training notebooks on Azure Databricks (PySpark, Delta Lake, Unity Catalog).
• Data Lake Architecture: Implement medallion (bronze/silver/gold) patterns on Azure Data Lake Storage Gen2; manage partitioning, file formats, and access patterns optimized for AI workloads.
• Data Integration: Develop ingestion and transformation pipelines using Azure Data Factory, Synapse, and Databricks Workflows to feed curated data into AI models and vector indexes.

AI Guardrails, Responsible AI & Security

• Guardrails Implementation: Implement input/output guardrails using Azure AI Content Safety, Prompt Shields, and groundedness checks; configure jailbreak, PII, and harmful-content filters at the model and gateway layers.
• Responsible AI: Build evaluation pipelines for safety, groundedness, relevance, and bias using Azure AI Foundry evaluations; embed Responsible AI checks into the deployment workflow.
• Security: Enforce private endpoints, VNet integration, Managed Identity, Key Vault, and RBAC across all AI services; ensure data residency and tenant isolation requirements are met.

Deployment Ownership & MLOps

• End-to-End Ownership: Take ownership of the full deployment lifecycle — from environment provisioning and CI/CD pipeline build to go-live, hypercare, and handover to operations.
• Infrastructure as Code: Author and maintain Terraform / Bicep modules for AI Foundry, Databricks, AML, Data Lake, AI Search, and supporting networking and identity components.
• CI/CD & MLOps: Build CI/CD pipelines (Azure DevOps or GitHub Actions) for prompt flows, model training, evaluation, and endpoint deployment; implement model registry, versioning, and promotion gates.
• Observability: Instrument AI workloads with Azure Monitor, Application Insights, and Log Analytics; set up dashboards for token usage, latency, cost, drift, and guardrail violations