support@Free-Braindumps.com
Register Login
Cisco
CompTIA
ISC
EC-Council
EXIN
Fortinet
ITIL
Juniper
Microsoft
VMware
Avaya
SAP
Google
NVIDIA
Salesforce
Amazon
Palo Alto Networks
ISC2
CrowdStrike
Anthropic
All Vendors
  1. Home
  2. Exams
  3. Google
  4. PROFESSIONAL MACHINE LEARNING ENGINEER

Free PROFESSIONAL MACHINE LEARNING ENGINEER Exam Braindumps
Professional Machine Learning Engineer Questions 36–40 (page: 9)

Updated On: 25-Oct-2026
Page 9 of 82
PDF with 402 Questions

You have developed an ML model to detect the sentiment of users’ posts on your company's social media page to identify outages or bugs. You are using Dataflow to provide real-time predictions on data ingested from Pub/Sub. You plan to have multiple training iterations for your model and keep the latest two versions live after every run. You want to split the traffic between the versions in an 80:20 ratio, with the newest model getting the majority of the traffic. You want to keep the pipeline as simple as possible, with minimal management required.
What should you do?

  1. Deploy the models to a Vertex AI endpoint using the traffic-split=0=80, PREVIOUS_MODEL_ID=20 configuration.
  2. Wrap the models inside an App Engine application using the --splits PREVIOUS_VERSION=0.2, NEW_VERSION=0.8 configuration
  3. Wrap the models inside a Cloud Run container using the REVISION1=20, REVISION2=80 revision configuration.
  4. Implement random splitting in Dataflow using beam.Partition() with a partition function calling a Vertex AI endpoint.

Answer(s): A



You developed an ML model with AI Platform, and you want to move it to production. You serve a few thousand queries per second and are experiencing latency issues. Incoming requests are served by a load balancer that distributes them across multiple Kubeflow CPU-only pods running on Google Kubernetes Engine (GKE). Your goal is to improve the serving latency without changing the underlying infrastructure.
What should you do?

  1. Significantly increase the max_batch_size TensorFlow Serving parameter.
  2. Switch to the tensorflow-model-server-universal version of TensorFlow Serving.
  3. Significantly increase the max_enqueued_batches TensorFlow Serving parameter.
  4. Recompile TensorFlow Serving using the source to support CPU-specific optimizations. Instruct GKE to choose an appropriate baseline minimum CPU platform for serving nodes.

Answer(s): A



You need to create a working environment in Agent Platform Workbench for a team of data scientists. Each data scientist has different VM CPU and RAM and package requirements, and will be assigned a personal notebook instance. You want each instance to have a custom set of packages pre-installed. Your company wants to minimize the running cost of notebook instances. How should you create the environment?

  1. Create Agent Platform Workbench instances based on a Deep Learning Containers container image. Disable idle shutdown of each instance.
  2. Create Agent Platform Workbench instances based on a Deep Learning Containers container image. Use n1-standard-4 as the default machine type. Verify idle shutdown of each instance.
  3. Create Agent Platform Workbench instances based on a custom container image derived from a Deep Learning Containers Image. Verify idle shutdown of each instance.
  4. Create Agent Platform Workbench instances based on a custom container image derived from a Deep Learning Containers image. Use n1-standard-4 as the default machine type. Verify idle shutdown of each instance.

Answer(s): C

Explanation:

A custom container image derived from a Deep Learning Containers image allows each notebook environment to include the required pre-installed packages while retaining a supported ML software base. Enabling idle shutdown reduces costs by automatically stopping notebook instances when they are not being used, while allowing each data scientist to select the VM CPU and RAM configuration appropriate for their workload.



You have a demand forecasting pipeline in production that uses Dataflow to preprocess raw data prior to model training and prediction. The only preprocessing step applies Z-score normalization on data stored in BigQuery and writes it back to BigQuery. New training data is added every week. You want to make the process more efficient by minimizing computation time and manual intervention.
What should you do?

  1. Normalize the data using Google Kubernetes Engine.
  2. Translate the normalization algorithm into SQL for use with BigQuery.
  3. Use the normalizer_fn argument in TensorFlow’s Feature Column API.
  4. Normalize the data with Apache Spark using the Dataproc connector for BigQuery.

Answer(s): B



You work at a large retail company that uses a time series forecasting model to predict daily sales. You have developed a new ARIMA_PLUS model to predict daily sales using historical sales data in BigQuery. The model has performed significantly better in offline evaluations than the current model which is deployed in a separate Agent Platform endpoint. You need to deploy the new model to production and replace the current production model. You want to manage the transition carefully to avoid any negative impact on business operations.
What should you do?

  1. Deploy both models to a new Agent Platform endpoint. Configure a 90/10 traffic split, sending 10% of the traffic to the new model, Monitor its performance, and gradually increase the traffic until it handles 100%.
  2. Deploy the new BigQuery ML model to a new Agent Platform endpoint. A/B test its performance against the old model using Agent Platform Experiments. If the new model performs better, update the production application to route all forecasting requests to the new endpoint.
  3. Deploy the new BigQuery ML model to a new Agent Platform endpoint. Update the production application to route all forecasting requests to the new endpoint.
  4. Run the new BigQuery ML model as a daily batch prediction job. Export the forecasts to a BigQuery table. Have the inventory management team compare these forecasts with the old model's forecasts for a few weeks before switching to the new model.

Answer(s): D

Explanation:

BigQuery ML ARIMA_PLUS models support time series forecasting through the ML.FORECAST function but cannot be deployed directly to Agent Platform endpoints. Running the new model as a daily batch forecasting job and exporting its predictions to a BigQuery table allows the company to evaluate its performance alongside the existing production model without disrupting business operations. Comparing forecasts over several weeks provides additional evidence of the new model's reliability under real-world conditions. This controlled, parallel evaluation approach minimizes deployment risk and supports a safe transition to the new forecasting solution.


Reference:

https://docs.cloud.google.com/bigquery/docs/reference/standard-sql/bigqueryml-syntax-create-mult ivariate-time-series https://docs.cloud.google.com/bigquery/docs/managing-models-vertex



  PDF
Viewing page 9 of 82
Viewing questions 41 - 45 out of 402 questions


Post your Comments and Discuss Google PROFESSIONAL MACHINE LEARNING ENGINEER exam prep with other Community members:

AI Tutor AI Tutor 👋 I’m here to help!