Home / ML Associate practice test / Model Deployment / Question

Databricks Machine Learning Associate · free practice question

A retailer needs nightly predictions for 40 million rows and can tolerate results arriving hours later. The team wants the cheapest and simplest approach…

Exam-style question on Model Deployment from the free ML Associate practice test, with the answer and a full explanation below.

Updated November 10, 2026

Model Deployment

A retailer needs nightly predictions for 40 million rows and can tolerate results arriving hours later. The team wants the cheapest and simplest approach rather than a low-latency API. Which serving pattern is the best fit?

  1. ABatch inference, because it is designed for large offline scoring jobs where low per-request latency is not required
  2. BReal-time inference, because every row should be sent individually to a REST endpoint to maximize responsiveness
  3. CStreaming inference, because a nightly workload is best handled as a continuous micro-batch service with checkpoints
  4. DInteractive notebook scoring, because manually running predictions each night provides the strongest production reliability
  5. ECanary deployment, because traffic splitting is the preferred way to score large historical tables in one pass
Answer & explanation

Correct answer: A

WHY A: Batch inference is the standard choice for large offline scoring workloads where latency is not critical and throughput matters most. WHY NOT B: Real-time serving adds unnecessary online API overhead. WHY NOT C: Streaming is aimed at continuously arriving data, not a simple nightly bulk run. WHY NOT D: Manual notebook execution is not the right production deployment pattern. WHY NOT E: Canary deployment is about rollout strategy, not bulk table scoring.

Take the full ML Associate practice test free →

Would you pass today?

This question is one of 45 in the free, timed Databricks Machine Learning Associate practice test. Take it and get a section-by-section score.

Start the free test →