Azure DP-100 study guide
Azure DP-100 · Work through your exam, one topic at a time.
Exam DP-100 leads to the Microsoft Certified: Azure Data Scientist Associate credential, an intermediate-level certification for data scientists who apply data science and machine learning to run workloads on Azure. Microsoft's certification page carries a warning that the certification and its renewal assessment are retired, so candidates should check Microsoft Learn for current credential options before scheduling. The role covers designing a data science working environment, exploring data, training models, implementing pipelines, running jobs for production, and managing, deploying, and monitoring scalable machine learning solutions. Candidates should also be able to use language models to build AI applications. The Azure DP-100 study guide has 4 sections and about 1,324 words.
01Design and Prepare a Machine Learning Solution
An Azure Machine Learning workspace is the top-level container. Creating one provisions (or attaches) an Azure Storage account - registered as the default datastore workspaceblobstore - a Key Vault that holds datastore secrets, Application Insights for monitoring, and an Azure Container Registry for environment images.
351 words, about 2 min02Explore Data and Run Experiments
An experiment is a logical grouping; each run is a job submitted with SDK v2 ( ml client.jobs.create or update ) or CLI v2 YAML. Track everything with MLflow : mlflow.log param , log metric , log artifact , plus autolog() for scikit-learn, PyTorch, and similar frameworks, which records parameters, metrics, and the model.
333 words, about 2 min03Train and Deploy Models
A command job bundles code , an environment , a compute target, and outputs . After training, register the artifact in the model registry with a type - mlflow model or custom model .
316 words, about 2 min04Optimize Language Models for AI Applications
Start with prompt engineering , add retrieval-augmented generation (RAG) when the model needs private or fresh knowledge, and use fine-tuning when you need consistent format, tone, or task behavior. RAG never changes model weights; fine-tuning is a poor way to inject facts that change often.
324 words, about 2 min