AIF-C01 Applications of Foundation Models Practice Question
A company uses a foundation model for real-time translation in a chat application. The latency is high. Which optimization would reduce latency the most?
⚠ Common exam trap
The AIF-C01 exam often tests the distinction between throughput optimization (batch size) and latency optimization (model size/distillation), leading candidates to mistakenly choose increasing batch size when the question explicitly asks for reducing latency.
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 model distillation to create a smaller model
Model distillation reduces the size of the foundation model by training a smaller 'student' model to mimic the behavior of a larger 'teacher' model. This directly decreases inference latency because the smaller model requires fewer computational resources (FLOPs) per forward pass, which is critical for real-time translation in a chat application where low latency is paramount.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Increase batch size
Why it's wrong here
Increasing batch size raises throughput by processing more requests per forward pass, but it also increases per-request latency, worsening real-time chat. It is tempting because batching is a standard efficiency technique, yet it optimises cost and throughput rather than the single-response latency this scenario demands.
- ✓
Use model distillation to create a smaller model
Why this is correct
Distillation trains a compact student model to mimic the larger teacher, cutting inference compute and memory per token. Fewer parameters mean faster forward passes, directly addressing the real-time chat latency constraint rather than merely tuning prompts or batching.
- ✗
Use a larger model
Why it's wrong here
A larger model adds parameters and computation per token, so inference latency rises rather than falls. It is tempting because larger models often improve translation quality, but quality is not the constraint here; the chat application needs lower response time, which smaller or distilled models deliver.
- ✗
Use a CDN for model weights
Why it's wrong here
A CDN caches static assets at edge locations; model weights are loaded once into the inference endpoint, not fetched per request, so a CDN does not reduce token-generation latency. It is tempting because CDNs speed up content delivery, but the bottleneck is model inference compute, not weight download.
Go deeper
Related to this question
About these practice questions
Courseiva writes every AIF-C01 question from scratch — 862 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or dumps. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This AIF-C01 practice question is part of Courseiva's free Amazon Web Services certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AIF-C01 exam.