easyMultiple Choice
PMLE Practice Question: A team has developed a prototype of a…
A team has developed a prototype of a recommendation model using a small dataset on a single VM. They need to scale to a larger dataset for production training. They plan to use Vertex AI training with a custom container. What is the best practice for handling the increased data volume?
⚠ Common exam trap
PMLE often tests the misconception that simply scaling up hardware (more memory, bigger batch) solves data volume issues, but the exam expects knowledge of efficient data formats and streaming for distributed training.
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 TFRecord format and streaming reads.
When scaling to a larger dataset for production training on Vertex AI with a custom container, the best practice is to use TFRecord format and streaming reads. TFRecord is a binary format optimized for TensorFlow, enabling efficient data serialization and streaming from Cloud Storage without loading the entire dataset into memory. This approach supports large-scale distributed training and reduces I/O bottlenecks.
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 the batch size to maximum.
Why it's wrong here
Increasing batch size raises memory per step and can degrade convergence; it does not address data volume exceeding a single machine. It is tempting because larger batches improve throughput on small datasets, and would be correct when GPU utilisation is low and memory headroom allows it.
- ✓
Use TFRecord format and streaming reads.
Why this is correct
TFRecord with streaming reads avoids loading the full dataset into memory, letting Vertex AI training workers read shards incrementally from Cloud Storage. This satisfies the increased data volume constraint, whereas downloading the entire dataset or using in-memory NumPy arrays would exhaust worker memory during distributed training.
- ✗
Store all data in memory before training.
Why it's wrong here
Loading the full dataset into memory cannot scale beyond the VM's RAM and defeats Vertex AI's distributed training. It is tempting because in-memory loading is fast and simple for small prototype datasets, and would be correct when the data fits comfortably in memory on a single machine.
- ✗
Use a single powerful VM with high memory.
Why it's wrong here
A single VM caps training at one machine's CPU, memory and accelerator limits, so it cannot scale with data volume. It is tempting because a large high-memory VM is an easy vertical upgrade from the prototype, and would be correct for workloads that genuinely cannot be distributed across multiple workers.
Go deeper
Related to this question
About these practice questions
One of 775 original PMLE practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official Google Cloud exam blueprint
This PMLE practice question is part of Courseiva's free Google Cloud 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 PMLE exam.