Microsoft AI-300 - Operationalizing Machine Learning and Generative AI Solutions (beta)

Imagine a hospital analytics firm with data scientists who kick off dozens of training jobs throughout the week. During peak hours, five jobs compete for the same GPU cluster and fail or queue for hours. On quiet nights, that cluster sits completely idle, burning money. Managed compute targets with autoscaling solve both problems: the cluster scales out automatically when multiple jobs arrive simultaneously and scales back to zero when idle. Option A (single shared cluster) is exactly the resource-contention problem Fabrikam already has. Option B (fixed-size cluster) wastes money during off-peak hours. Option C (dedicated per-experiment clusters) eliminates contention but is prohibitively expensive for a cost-conscious healthcare firm. Autoscaling managed compute is the cloud-native answer to variable workload demand.

Microsoft Learn Reference Topic: Azure Machine Learning compute targets – Configure autoscale for compute clusters

MLflow is the industry-standard open-source platform for experiment tracking, and Azure Machine Learning has first-class native integration with it. When you use MLflow within an Azure ML job, parameters, metrics, and artifacts are automatically logged to the run history of the AML workspace, making every run reproducible and comparable. Option A (AML job output logs) only captures console output and lacks structured parameter and metric logging. Option C (Application Insights logs) is designed for application-level telemetry, not ML experiment metadata. Option D (Azure Monitor alerts) is a reactive notification tool, not a tracking system. MLflow ' s autologging capability means that for common frameworks such as scikit-learn, XGBoost, and PyTorch, parameters and metrics are captured without a single line of custom code, directly answering Fabrikam ' s requirement for consistent experiment tracking.

Microsoft Learn Reference Topic: Track ML experiments with MLflow – Azure Machine Learning MLflow integration

Managed Identity is Azure ' s solution for eliminating stored secrets and API keys from application configurations. When a Microsoft Foundry resource or its dependent compute has a system-assigned or user-assigned managed identity, it can authenticate to Azure services such as Storage, Key Vault, and Cognitive Services using Azure Entra ID tokens that are automatically issued and rotated by Azure. There are no keys to store, rotate, or accidentally expose in code or configuration files, directly satisfying the requirement to authenticate without stored credentials. Role-Based Access Control (RBAC) is the Azure governance mechanism for scoping permissions. By assigning specific built-in roles to the managed identity or to user and group principals at the correct scope — subscription, resource group, resource, or project level — the organization can enforce least-privilege access across different projects and shared resources.

Microsoft Learn Reference Topic: Configure managed identities and RBAC for Microsoft Azure AI Foundry resources

Enabling tracing via the prompt flow SDK alone does not guarantee that all four required dimensions — inputs, outputs, token usage, and latencies — are captured and surfaced in the Microsoft Foundry portal ' s trace viewer in a way that allows manual comparison across runs. While the prompt flow SDK does support tracing configuration, the portal-based Tracing activation is the complete, supported path for this specific capture requirement. The correct and complete solution as confirmed by the next question in this series is to use the Tracing toggle directly in the Microsoft Foundry portal for the project, then execute test runs. This portal-side Tracing feature is purpose-built to instrument all four dimensions and store them in a queryable trace store accessible through the Foundry interface.

Microsoft Learn Reference Topic: Enable tracing for prompt flows in Microsoft Foundry portal – Capturing run telemetry

Azure Machine Learning infrastructure can be described as Infrastructure as Code using ARM templates, Bicep, or Terraform. To make provisioning fully automated, two separate concerns must be addressed. First, the trigger: a CI/CD system like GitHub Actions or Azure DevOps Pipelines must detect changes to IaC template files in the repository, configured through a branch-push or pull-request trigger. Second, the execution: the CI/CD pipeline runs the provisioning commands such as az ml workspace create or bicep deploy non-interactively using a service principal or managed identity, eliminating the need for a human to run scripts manually. Together, these two mechanisms deliver the GitOps-style automation that modern MLOps practices require, ensuring every template change automatically results in a consistent, auditable infrastructure update.

Microsoft Learn Reference Topic: Automate Azure Machine Learning with GitHub Actions – CI/CD for ML infrastructure

GPT-4o is OpenAI ' s multimodal model that accepts both text and image inputs natively. The ' o ' in GPT-4o stands for ' omni, ' reflecting its ability to process multiple modalities in a single inference call. It is available in Microsoft Foundry for both inference and fine-tuning, and it supports vision inputs alongside text prompts, making it the correct choice for any fine-tuning scenario requiring multimodal input. Davinci-002 (option A) is a legacy text-completion model with no vision capabilities and limited fine-tuning support for modern chat formats. GPT-3.5-Turbo (option C) is a text-only chat model that does not accept image inputs. GPT-4 (option D) has a standard version that is text-only; the vision variant GPT-4V is distinct from GPT-4o and has more limited fine-tuning availability in Foundry. For multimodal text plus image fine-tuning, GPT-4o is the only correct choice.

Microsoft Learn Reference Topic: Azure OpenAI Service models available for fine-tuning – GPT-4o multimodal capabilities

A loan approval model must be assessed for both transparency and demographic fairness before registration. Option D (feature importance) addresses prediction transparency — using SHAP or Azure ML ' s Responsible AI dashboard, the team can see which features drive individual predictions, revealing whether proxy variables for protected attributes are influencing decisions inappropriately. Option E (error rates across demographic cohorts) directly measures whether the model makes more mistakes for specific groups — a core fairness metric required under Responsible AI principles. Option A (global error rates) masks disparities since you need cohort-level analysis. Option B (endpoint latency) is a performance engineering concern. Option C (inference schema compatibility) is a technical integration check. The Azure ML Responsible AI dashboard combines both error analysis and feature importance in a unified interface designed for exactly this type of pre-deployment fairness assessment.

Microsoft Learn Reference Topic: Responsible AI dashboard in Azure Machine Learning – Error analysis and model interpretability

During high traffic, the question is whether the endpoint is keeping up with demand. Requests per minute (B) tells you the actual request load on the endpoint, letting you correlate traffic spikes with degradation. Connections active (C) reveals whether the endpoint ' s connection pool is saturating — too many concurrent connections without adequate scaling causes queuing and timeouts. Request latency (E) is the gold-standard measure of user-perceived performance; rising latency under load is the clearest signal of degradation. Feature count (A) is a model-design attribute, not a runtime performance metric. Dataset size (D) is a training-time concern unrelated to endpoint performance. Azure Machine Learning online endpoints expose these metrics through Azure Monitor, and Microsoft recommends configuring alert rules on latency and request rate thresholds for all production endpoints.

Microsoft Learn Reference Topic: Monitor Azure Machine Learning online endpoints – Azure Monitor metrics for managed endpoints