Databricks Machine Learning Professional Exam - Topic 8 Question 19 Discussion

Actual exam question for Databricks's Databricks Machine Learning Professional exam

Question #: 19
Topic #: 8

[All Databricks Machine Learning Professional Questions]

A machine learning engineer needs to select a deployment strategy for a new machine learning application. The feature values are not available until the time of delivery, and results are needed exceedingly fast for one record at a time.

Which of the following deployment strategies can be used to meet these requirements?

AEdge/on-device

BStreaming

CNone of these strategies will meet the requirements.

DBatch

EReal-time

Show Suggested Answer

Suggested Answer: B

by Alva at Jul 03, 2024, 08:45 PM

Limited Time Offer

25%

Off

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

Contribute your Thoughts:

Submit Cancel

Jacqueline

7 months ago

Streaming could be a good fit too, right?

upvoted 0 times

...

Lino

8 months ago

Definitely not batch, that's way too slow!

upvoted 0 times

...

Jacquelyne

8 months ago

Wait, can none of these strategies actually work?

upvoted 0 times

...

Corazon

8 months ago

I think real-time is the way to go here.

upvoted 0 times

...

Trina

8 months ago

Edge/on-device is a solid choice for fast results!

upvoted 0 times

...

Cory

9 months ago

Batch processing seems too slow for this requirement, so I guess it's between edge/on-device and real-time, but I’m leaning towards real-time.

upvoted 0 times

...

Elza

9 months ago

I think I came across a question similar to this in practice, and it mentioned that streaming could be useful for continuous data, but I'm not convinced it fits this scenario.

upvoted 0 times

...

Samuel

9 months ago

I'm not entirely sure, but I feel like real-time processing could also work here since it emphasizes speed for individual records.

upvoted 0 times

...

Carmelina

9 months ago

I remember studying deployment strategies, and I think edge/on-device might be a good fit since it can process data locally without waiting for a server response.

upvoted 0 times

...

Timothy

9 months ago

I'm a little confused by this question. None of the options seem like a perfect fit to me. The lack of feature values until delivery is a tricky constraint. I'm leaning towards the edge/on-device or streaming strategies, but I'm not 100% sure. Guess I'll have to make an educated guess on this one.

upvoted 0 times

...

Malcom

9 months ago

Hmm, I'm a bit unsure about this one. The lack of feature values until delivery is throwing me off. I'm not sure if batch or real-time would work since they might be too slow. I'll have to think this through carefully.

upvoted 0 times

...

Ranee

9 months ago

This seems like a straightforward question. The key requirements are that the feature values aren't available until delivery and the results need to be extremely fast for one record at a time. I think the edge/on-device or streaming options are the most likely to meet those needs.

upvoted 0 times

...

Anabel

9 months ago

Okay, I think I've got it. The edge/on-device option seems like the best fit since the model can be deployed directly on the device and provide fast results for individual records without needing the feature values upfront. Streaming could also work, but might be overkill for this use case.

upvoted 0 times

...

Sabra

9 months ago

Wait, I'm confused. There are a few similar-looking options here, and I'm not totally sure which one is correct. I'll have to eliminate the ones that don't seem quite right and then make my best guess.

upvoted 0 times

...

Myong

9 months ago

Hmm, I'm a bit unsure about this one. The options seem similar, and I want to make sure I understand the differences between them before selecting an answer.

upvoted 0 times

...

Almeta

10 months ago

I'm feeling pretty confident about this one. The concepts of use case development and scenario types are something I'm familiar with, so I should be able to spot the false statement.

upvoted 0 times

...

Deonna

10 months ago

I feel pretty confident about this one. The outcome has to be the consequence or result of the event, not the event itself. I'll go with A.

upvoted 0 times

...

Apolonia

1 year ago

Batch deployment is definitely not the way to go here. Who needs slow and outdated results when you can have lightning-fast responses? Edge/on-device all the way!

upvoted 0 times

Macy

1 year ago

Streaming could be another option for fast delivery of results.

upvoted 0 times

...

Clarinda

1 year ago

Real-time deployment could also work for quick responses.

upvoted 0 times

...

Nidia

1 year ago

Edge/on-device is the way to go for fast results!

upvoted 0 times

...

Katheryn

1 year ago

I'm not sure any of these strategies will fully meet the requirements. Maybe they should consider a quantum computer instead? Just a thought...

upvoted 0 times

...

Janella

1 year ago

Streaming deployment could also work, as it can handle real-time data processing. But it might not be as fast as edge/on-device for a single record.

upvoted 0 times

Maryann

1 year ago

C: None of these strategies will meet the requirements.

upvoted 0 times

...

Hildred

1 year ago

B: Real-time

upvoted 0 times

...

Scarlet

1 year ago

A: Edge/on-device

upvoted 0 times

...

Venita

1 year ago

I'm not sure, but I think A) Edge/on-device could also work since the feature values are not available until delivery.

upvoted 0 times

...

Shayne

1 year ago

I agree with Amie, Real-time deployment strategy would be the most suitable for this scenario.

upvoted 0 times

...

Vivan

1 year ago

Edge/on-device deployment seems like the best option here since the feature values are only available at the time of delivery and the results are needed very quickly for individual records.

upvoted 0 times