A team deploys a containerized model on Kubernetes. They need to ensure that the model can handle a traffic spike without downtime, but they also want to avoid over-provisioning resources during normal periods. The model is stateless. Which approach BEST meets these requirements?

Use a CronJob to run the model periodically

Use a single replica and rely on the Service to handle traffic

Use a Deployment with multiple replicas and a HorizontalPodAutoscaler based on CPU utilization

Use a StatefulSet with persistent volumes

A team deploys a model on AWS SageMaker. They need to integrate the endpoint with an application that sends requests from a VPC. The endpoint should only be accessible from within the VPC. What should they do?

Place the endpoint behind a NAT gateway

Configure the SageMaker endpoint to be private and use VPC mode

Use SageMaker Ground Truth to label the endpoint

Use a public endpoint and restrict access with IAM policies

A data scientist has a trained model saved as a pickle file. They need to package it into a Docker image so it can be served as a REST API. What is the MINIMUM set of steps to accomplish this?

Use Docker Compose to orchestrate the model and database

Write a Dockerfile that installs dependencies, copies the model and a serving script, then build the image and run the container

Copy the pickle file into a base Python image and run it

Push the pickle file to Docker Hub

A data science team uses AWS SageMaker to deploy a model. They need to test a new model version with a small percentage of live traffic before rolling it out fully. Which SageMaker feature should they use?

Use SageMaker Ground Truth to label the new model's outputs

Create a new endpoint configuration with multiple production variants and set traffic weights

Deploy the new version to a separate endpoint and manually route traffic

Create a SageMaker Processing job to evaluate the new model

CertNexusCertified Data Science Practitioner (CDSP)

Domain 6Objective 2

Deploy and Manage Models CERTIFIED-DATA-SCIENCE-PRACTITIONER Practice Questions (Page 3)

Part of the 6.0 Operationalizing the pipeline domain, which accounts for 5-8% of the CERTIFIED-DATA-SCIENCE-PRACTITIONER exam.

36questions here

8free pages

8concepts

5-8%of the exam

Questions 11–15

11foundation · easy
A model deployed for real-time fraud detection must return a prediction in under 100 milliseconds. Which production environment consideration is most directly addressed by this requirement?
Select an answer first
12expert · hard
A team deploys a containerized model on Kubernetes. They need to ensure that the model can handle a traffic spike without downtime, but they also want to avoid over-provisioning resources during normal periods. The model is stateless. Which approach BEST meets these requirements?
Select an answer first
13application · medium
A team deploys a model on AWS SageMaker. They need to integrate the endpoint with an application that sends requests from a VPC. The endpoint should only be accessible from within the VPC. What should they do?
Select an answer first
14application · medium
A data scientist has a trained model saved as a pickle file. They need to package it into a Docker image so it can be served as a REST API. What is the MINIMUM set of steps to accomplish this?
Select an answer first
15application · medium
A data science team uses AWS SageMaker to deploy a model. They need to test a new model version with a small percentage of live traffic before rolling it out fully. Which SageMaker feature should they use?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to CERTIFIED-DATA-SCIENCE-PRACTITIONER

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by CertNexus. “CERTIFIED-DATA-SCIENCE-PRACTITIONER” is a trademark of its owner, used for identification only.