Google Professional Machine Learning Engineer Question 160
Single answerGoogle Cloud PlatformYour organization is using Dataproc on Google Cloud to process large-scale data with Spark. You notice that some machine learning workloads are running inefficiently due to the lack of interactivity during data exploration and experimentation. How can you improve the workflow using Spark kernels to optimize interactivity while leveraging Dataproc?
- A
Use Jupyter Notebooks integrated with Dataproc to run Spark kernels for interactive data exploration.
- B
Switch to using TensorFlow kernels instead of Spark kernels for better performance in Dataproc.
- C
Deploy Cloud Functions to handle Spark jobs in Dataproc and improve interactivity.
- D
Enable Dataproc's Jupyter component to create a managed environment for running Spark kernels interactively.
Show answer and explanation
Correct answer: A
Explanation
To optimize interactivity during data exploration and experimentation on Dataproc, Spark kernels should be used in an interactive environment like Jupyter Notebooks. Dataproc supports integration with Jupyter, making it an ideal setup for running Spark workloads interactively. This approach ensures efficient use of resources and improved productivity during machine learning workflows.
- A. Correct.
Using Jupyter Notebooks with Spark kernels allows you to leverage interactive data exploration and experimentation directly in Dataproc. This is the most efficient and appropriate solution for the described scenario.
- B. Incorrect.
TensorFlow kernels are not designed for Spark-based data processing and cannot improve interactivity in Spark workloads. This is not relevant to the problem described.
- C. Incorrect.
Cloud Functions are event-driven and not designed for interactive data exploration or running Spark kernels. They are more suited for lightweight, stateless tasks.
- D. Incorrect.
While enabling the Jupyter component in Dataproc is beneficial, it is only part of the solution. You need to specifically configure and use Spark kernels for interactivity, which is not explicitly addressed by this option.