AI-300 case studies
AI-300 · Put your knowledge to work in context.
Exam AI-300 is the requirement for the Microsoft Certified: Machine Learning Operations Engineer Associate credential. It targets engineers who build and operate AI workloads on Azure across two tracks: classic MLOps using Azure Machine Learning, and GenAIOps using Microsoft Foundry. Microsoft expects candidates to have a data science background with Python, entry-level DevOps familiarity (GitHub Actions, command-line interfaces), and hands-on experience with infrastructure as code using Bicep and Azure CLI. The exam measures five areas: designing and implementing MLOps infrastructure (15–20%), machine learning model lifecycle and operations (25–30%), GenAIOps infrastructure (20–25%), generative AI quality assurance and observability (10–15%), and optimizing generative AI systems and model performance (10–15%). The AI-300 case studies are 5 cases and about 3,360 words.
1Consolidating a Fragmented Azure ML Estate Without Losing Lineage
A financial-services company runs 45 data scientists across 6 teams on Azure Machine Learning. Each team owns its own workspace, model registry, and AKS inference cluster; the same feature-engineering code lives in four repositories.
682 words, about 3 min2Retraining Triggers for a Demand Model With Week-Delayed Labels
A retailer forecasts demand for 400 SKUs with a model retrained monthly regardless of need. Ground-truth sales arrive about 7 days late.
690 words, about 3 min3Prompt and Index Versioning for a RAG Assistant in CI/CD
A RAG assistant answers questions over 120,000 internal documents. Prompt edits are made directly in the Azure portal by three engineers and one product manager.
636 words, about 3 min4Calibrating an LLM Judge for a Support Copilot
A support copilot handles 60,000 conversations a day. Quality is measured by an LLM judge on 100 sampled conversations per day, scoring groundedness and helpfulness.
656 words, about 3 min5Cutting p95 Latency and Cost on a Mixed-Traffic LLM Application
An internal assistant serves 500,000 requests a day against a large frontier model. p95 latency is 9.2s against a 3s SLA, and cost is $0.04 per request.
696 words, about 3 min