PMLE Scaling Prototypes into ML Models Practice Question

You are training a scikit-learn random forest on a dataset that fits in memory using Vertex AI custom training. The prototype notebook took 15 minutes, but the Vertex AI job takes over an hour and occasionally fails with a resource error. You want the production job to complete reliably without changing the model or preprocessing. What should you do?

⚠ Common exam trap

The trap here is assuming that any training job that fails must be fixed with distributed training, when a single-machine resource increase is the correct and simpler remedy.

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

✓

Select a machine type with more memory and vCPUs, and set the appropriate boot disk size for the training job.

The job fits in memory but fails on the default Vertex AI worker configuration, which points to insufficient RAM or vCPUs on the training machine. Choosing a larger machine type and an adequate boot disk gives the scikit-learn process the resources it needs. Distributed training is unnecessary and would require changing the algorithm, while preemptible VMs reduce reliability rather than improve it.

Answer analysis

Option-by-option breakdown

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

✗
Switch the training job to use a preemptible VM to reduce contention on shared resources.
Why it's wrong here
Preemptible VMs are lower-cost but can be reclaimed at any time, which makes a job less reliable, not more. They are appropriate for fault-tolerant workloads with checkpointing, not for a job that currently fails with a resource error. Using them here would increase the chance of interruption and would not address the memory or CPU shortage causing the failure.
✓
Select a machine type with more memory and vCPUs, and set the appropriate boot disk size for the training job.
Why this is correct
Vertex AI custom training lets you choose a machine type and boot disk for the worker pool. A scikit-learn job that fits in memory but fails on a small default machine benefits from more RAM and vCPUs. Increasing available memory prevents out-of-memory failures, and a larger boot disk ensures the container and any temporary files have sufficient space, matching the reliability goal without changing the model.
✗
Configure the training job to run on a single n1-standard-4 machine with no accelerator.
Why it's wrong here
A single n1-standard-4 provides only 15 GB of RAM and four vCPUs. If the dataset or model memory footprint exceeds that, the job can fail with an out-of-memory or resource-exhausted error. Right-sizing to a larger machine type, not a smaller one, addresses the reliability problem described, so this choice would likely worsen the failure rather than fix it.
✗
Increase the number of worker replicas and use a distributed reduction strategy for the random forest.
Why it's wrong here
scikit-learn's RandomForestRegressor and RandomForestClassifier do not provide native multi-node distributed reduction. Adding worker replicas would require a different framework or a custom distributed algorithm, which changes the model and preprocessing. The scenario explicitly says the model should not change, so distributed training is not the appropriate fix for this single-machine resource issue.

Go deeper

PMLE Scaling Prototypes into ML Models Practice Question

Why each option matters

Option-by-option breakdown

Related to this question