Endpoint Selection & Real Time Inference — MLA-C01 Practice Question
A representative AWS Machine Learning Engineer Associate (MLA-C01) exam question on Endpoint Selection & Real Time Inference. Work through it below, then read why each option is right or wrong.
Short answer
The correct answer is B. A SageMaker real-time (persistent) endpoint.
A persistent real-time endpoint keeps instances warm behind a low-latency HTTPS endpoint, which is exactly what's needed for sub-second scoring during a live checkout flow with sustained, predictable traffic.
The Question
A ride-sharing company built a fraud-detection model that must score each payment transaction within 100 ms while the app is processing the checkout flow, with steady traffic of roughly 500 requests per second throughout the day. Which SageMaker inference option meets these requirements?
Why B is correct
A persistent real-time endpoint keeps instances warm behind a low-latency HTTPS endpoint, which is exactly what's needed for sub-second scoring during a live checkout flow with sustained, predictable traffic.
Why the other options are wrong
Batch Transform processes a stored dataset offline in a job and returns results only after the whole job finishes; it cannot satisfy a synchronous 100 ms request.
Asynchronous Inference is designed for large payloads or long-running inference (minutes), queuing requests and returning results later — not for interactive low-latency scoring.
Serverless Inference is a good fit for intermittent or unpredictable traffic with cold-start tolerance, but steady high-volume traffic (500 TPS) is better and more predictably served by a provisioned real-time endpoint, since serverless has cold-start latency risk and concurrency limits.
Key idea: Endpoint Selection & Real Time Inference
Why B is correct: A persistent real-time endpoint keeps instances warm behind a low-latency HTTPS endpoint, which is exactly what's needed for sub-second scoring during a live checkout flow with sustained, predictable traffic. Why A is wrong: Batch Transform processes a stored dataset offline in a job and returns results only after the whole job finishes; it cannot satisfy a synchronous 100 ms request. Why C is wrong: Asynchronous Inference is designed for large payloads or long-running inference (minutes), queuing requests and returning results later — not for interactive low-latency scoring. Why D is wrong: Serverless Inference is a good fit for intermittent or unpredictable traffic with cold-start tolerance, but steady high-volume traffic (500 TPS) is better and more predictably served by a provisioned real-time endpoint, since serverless has cold-start latency risk and concurrency limits. On the MLA-C01 exam, questions in the "Deployment and Orchestration of ML Workflows" domain test whether you can map a scenario's constraints to the right choice. Read the requirement carefully, eliminate options that violate any single constraint, and pick the one that satisfies all of them with the least operational overhead.
Ready to see how you'd score?
Take the free practice quiz and find out which AWS Certified Machine Learning Engineer - Associate domains you need to focus on. No signup required.
Practice 5 similar questions
Same cert, same or adjacent domain. Use these after reviewing the explanation.