AIF-C01 Fundamentals of AI and ML Practice Question

A financial institution is deploying a fraud detection model using Amazon SageMaker. The model must be able to handle sudden spikes in inference requests during promotional events while keeping costs low. The team wants to use a serverless architecture to avoid provisioning idle capacity and to scale automatically from zero. However, the inference latency requirement is under 5 seconds for each request. Which SageMaker inference option should they choose?

⚠ Common exam trap

AWS often tests the misconception that serverless inference cannot meet low-latency requirements, but SageMaker Serverless Inference is specifically designed for sub-second to few-second latency, making it suitable for real-time fraud detection scenarios.

Answer choices

Why each option matters

Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.

Correct answer & explanation

✓

Use Amazon SageMaker Serverless Inference

Amazon SageMaker Serverless Inference is the correct choice because it automatically scales from zero to handle sudden spikes in inference requests, aligning with the requirement to avoid provisioning idle capacity. It also meets the sub-5-second latency requirement for fraud detection, as it is designed for low-latency, on-demand inference without managing underlying infrastructure.

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

✓
Use Amazon SageMaker Serverless Inference
Why this is correct
Serverless Inference provisions compute on demand, scales automatically from zero and charges only per request, eliminating idle capacity costs. It meets the under-five-second latency requirement, since cold-start and invocation latency stay within that threshold for typical payloads.
✗
Use Amazon SageMaker Multi-Model Endpoints
Why it's wrong here
Multi-Model Endpoints host many models behind one endpoint, sharing instances by loading models on demand; they scale with instance count and cannot scale to zero, so idle capacity remains billable. They suit hosting thousands of low-traffic models, not a single fraud model needing serverless zero-to-scale behaviour.
✗
Use Amazon SageMaker real-time endpoints with auto-scaling
Why it's wrong here
Real-time endpoints run continuously on provisioned instances, so they cannot scale to zero and idle capacity is billed between promotional spikes. Auto-scaling adjusts instance counts reactively, not from zero. They suit steady, latency-critical traffic, not the serverless cost model this scenario demands.
✗
Use Amazon SageMaker Asynchronous Inference
Why it's wrong here
Asynchronous Inference queues requests and returns results via Amazon S3, designed for payloads up to 1 GB and processing times up to 15 minutes; it does not return a synchronous response within 5 seconds. It suits long-running, large-payload jobs, not interactive fraud scoring with sub-5-second latency.

Quick reference

Cloud Service Model Comparison

Model	You Manage	Provider Manages	Examples
IaaS	OS, runtime, apps, data	Hardware, hypervisor, networking	EC2, Azure VMs, GCP Compute Engine
PaaS	Apps and data	OS, runtime, middleware, hardware	Elastic Beanstalk, Azure App Service
SaaS	Data and settings only	Everything else	Microsoft 365, Salesforce, Workday
FaaS / Serverless	Function code only	Infra, scaling, runtime	Lambda, Azure Functions, Cloud Run
CaaS	Containers and apps	Kubernetes, OS, hardware	EKS, AKS, GKE

Go deeper

AIF-C01 Fundamentals of AI and ML Practice Question

Why each option matters

Option-by-option breakdown

Cloud Service Model Comparison

Related to this question