Azure ML Batch Endpoint Deployment Order

DP-100 By islever Editorial Last updated

Short answer

A scoring script

Question

You manage an Azure Machine Learning workspace. An MLflow model is already registered. You plan to customize how the deployment does inference. You need to deploy the MLflow model to a batch endpoint for batch inferencing. What should you create first?
  1. scoring script Correct Answer
  2. deployment
  3. environment
  4. deployment definition

Key Takeaways

The exam tests the logical dependency chain of deployment artifacts; the common trap is confusing the order between creating the definition (metadata) versus the actual execution code (scoring script).

This question tests the prerequisite components for deploying an MLflow model to a batch endpoint in Azure Machine Learning, with community consensus favoring the scoring script as the initial creation step.

Expert Analysis

Why the Answer Is Correct

In the context of Azure Machine Learning batch deployments, particularly when using MLflow models, the scoring script is the fundamental artifact that defines how the model performs inference. While the deployment definition acts as a template or configuration object, the scoring script contains the actual Python logic required to load the model and process input data. Without a valid scoring script, the deployment cannot function, making it the primary component to establish first in the workflow.

Why the Other Options Are Wrong

A deployment (Option B) cannot be created without referencing both an environment and a scoring script, so it is not the first step. An environment (Option C) is necessary for dependencies but does not contain the inference logic itself. A deployment definition (Option D) is a configuration file or object that references the scoring script and environment; while you might define the structure early, the scoring script is the core executable asset that must exist to make the deployment viable.

Community Comment Notes

Community comments show a split opinion, with some arguing for the deployment definition (D) based on infrastructure-as-code workflows, while others (including the suggested answer) emphasize the scoring script (A) as the functional starting point. Comment [1] highlights that initialization requires the run logic found in the scoring script. Comment [2] argues for D, suggesting scripts can be added later, which contradicts the strict dependency where the definition points to the script. The majority vote (67%) supports A, indicating this is the expected 'best practice' answer for this specific exam scenario.

Exam Strategy

When asked about deployment prerequisites, prioritize the artifact that contains the actual business logic or inference code (like the scoring script) over the metadata containers (like definitions or environments).

Community Votes

A
67%
D
33%

67% of anonymous learners picked answer A. Votes are pick records left by other test-takers — they are not the verified answer.

Discussion 3 contributions

Compiled from public community threads and reader submissions, lightly edited for clarity. Handles and vote tallies are shown as recorded, and comments reflect the views of individual contributors.

From public community threads 3

  1. jl420 👍 1 Answered D
    D is correct. Scoring Script can be done later.
  2. Shudharsanan 👍 1 Answered A
    init and run. so scoring script
  3. onurag 👍 1 Answered A
    Should be scoring_script

Reader comments & corrections

No reader corrections yet — spotted an error or have a note? Share it below.

Log in to comment, report an error, or add a note about this question.

Azure ML Workspace Authentication with MLClient.from_config

S-grade

Deploy MLflow Model to Endpoint Without Egress Connectivity

S-grade

Azure ML Hyperparameter Tuning: Random Sampling vs. Early Termination

S-grade

Update Synapse Spark Compute Identity in Azure ML

S-grade

← Back to DP-100 Study Guide