Google Professional Machine Learning Engineer Exam - Topic 5 Question 4 Discussion

Actual exam question for Google's Professional Machine Learning Engineer exam

Question #: 4
Topic #: 5

[All Professional Machine Learning Engineer Questions]

You work with a data engineering team that has developed a pipeline to clean your dataset and save it in a Cloud Storage bucket. You have created an ML model and want to use the data to refresh your model as soon as new data is available. As part of your CI/CD workflow, you want to automatically run a Kubeflow Pipelines training job on Google Kubernetes Engine (GKE). How should you architect this workflow?

AConfigure your pipeline with Dataflow, which saves the files in Cloud Storage After the file is saved, start the training job on a GKE cluster

BUse App Engine to create a lightweight python client that continuously polls Cloud Storage for new files As soon as a file arrives, initiate the training job

CConfigure a Cloud Storage trigger to send a message to a Pub/Sub topic when a new file is available in a storage bucket. Use a Pub/Sub-triggered Cloud Function to start the training job on a GKE cluster

DUse Cloud Scheduler to schedule jobs at a regular interval. For the first step of the job. check the timestamp of objects in your Cloud Storage bucket If there are no new files since the last run, abort the job.

Show Suggested Answer

Suggested Answer: A

by Carmen at May 05, 2022, 02:41 PM

Limited Time Offer

25%

Off

Get Premium Professional Machine Learning Engineer Questions as Interactive Web-Based Practice Test or PDF

Contribute your Thoughts:

Submit Cancel

Bettina

4 months ago

Option D seems outdated; polling isn't the best practice anymore.

upvoted 0 times

...

Zachary

4 months ago

I like option B for its real-time approach, but it might be resource-heavy.

upvoted 0 times

...

Aileen

4 months ago

Wait, can you really use Cloud Functions like that? Sounds too easy!

upvoted 0 times

...

Thurman

4 months ago

I disagree, I think option A is simpler and more straightforward.

upvoted 0 times

...

Ceola

5 months ago

Option C sounds like the most efficient way to trigger the training job!

upvoted 0 times

...

Yen

5 months ago

I’m leaning towards option C as well. The idea of using a Cloud Storage trigger seems efficient, but I wonder if there are any latency issues with the Cloud Function starting the job on GKE.

upvoted 0 times

...

Virgilio

5 months ago

I think using App Engine to poll for new files, like in option B, could work, but it feels a bit inefficient. I recall we practiced a similar question where polling wasn't the best approach.

upvoted 0 times

...

Alex

5 months ago

I remember we discussed using Pub/Sub for event-driven architectures in our last class. Option C seems to align with that concept, but I'm not entirely sure about the Cloud Function part.

upvoted 0 times

...

Loreta

5 months ago

Option A sounds familiar, but I can't quite recall if Dataflow is the right tool for this specific scenario. I feel like there might be a better way to trigger the training job.

upvoted 0 times

...

Bernadine

5 months ago

Hmm, I'm not sure about this one. I can see how geolocation would be important, but I'm also wondering if the camera could be useful for things like taking photos of landmarks. I'll have to think this through a bit more.

upvoted 0 times

...

Skye

5 months ago

Okay, I think I've got a handle on this. The Simplicity 380 and Bladesystem 7000 both sound like they could work, but the Converged systems option seems to best match the requirements of minimizing footprint and quick deployment. I'll go with that.

upvoted 0 times

...

Leota

5 months ago

I remember practicing a question like this where the focus was on gathering comprehensive information, which might lean towards option C here.

upvoted 0 times

...