Practice for the Red Hat Certified Specialist in OpenShift AI (EX267) exam with 17 exam-style practice questions, instant answer reveals, and concise explanations of every correct answer. Topics include: Which method permits passing an API authentication token to securely access an auth-protected KServe endpoint via curl?. Follow @CertPunch and visit certpunch.com for more certification practice exams and study content.
Prefer hands-on? Take this round as an interactive practice test — answer every question, get instant feedback, and see your score: Start the Red Hat Certified Specialist in OpenShift AI (EX267) practice test →
What you will practice
- Which method permits passing an API authentication token to securely access an auth-protected KServe endpoint…
- Which model serving runtime is natively integrated into OpenShift AI for high-performance LLM and structured…
- Which interface option allows an engineer to clone a remote Git repository securely within the JupyterLab das…
- Which specific profile allows platform administrators to introduce custom accelerators like Intel Gaudi or AM…
- WHAT TYPE OF CLUSTER ENVIRONMENT OBJECT HANDLES PIPELINE METADATA STORAGE WITHIN AN OPENSHIFT AI ENVIRONMENT?
- Which API object configuration captures Prometheus scraping endpoints, attaching comprehensive AI endpoint re…
Answers and explanations
Tap a question to expand the answer and the exam reasoning. Try to commit to your own pick first.
Q1. Which method permits passing an API authentication token to securely access an auth-protected KServe endpoint via curl?
Answer: C. Passing the bearer token in the HTTP Authorization header (-H 'Authorization: Bearer ')
Passing the bearer token in the HTTP Authorization header is the standard and secure way to authenticate to a protected KServe endpoint. Remember that placing tokens in the URL or payload is insecure, and disabling SSL verification exposes the session to interception.
Q2. Which model serving runtime is natively integrated into OpenShift AI for high-performance LLM and structured model deployment?
Answer: B. Caikit-TGIS Runtime
The Caikit-TGIS Runtime is specifically integrated to provide high-performance serving for large language models within OpenShift AI. Traditional web servers like Apache HTTPD or application servers lack the specialized inference capabilities required for natural language processing tasks.
Q3. Which interface option allows an engineer to clone a remote Git repository securely within the JupyterLab dashboard?
Answer: D. Git extension in the sidebar using SSH or HTTPS authentication
The built-in Git extension in the JupyterLab sidebar provides a graphical interface for securely cloning repositories using standard authentication. While you could technically use a terminal shell, the dashboard option specifically refers to this integrated Git extension.
Q4. Which specific profile allows platform administrators to introduce custom accelerators like Intel Gaudi or AMD Instinct drivers uniformly?
Answer: A. AcceleratorProfiles Custom Resource objects in the OpenShift AI management space
AcceleratorProfiles custom resources define how OpenShift AI identifies and utilizes specialized hardware accelerators across the cluster. Modifying base operating system configurations or node labels lacks the centralized abstraction needed to properly map these runtime resources to workbenches.
Q5. WHAT TYPE OF CLUSTER ENVIRONMENT OBJECT HANDLES PIPELINE METADATA STORAGE WITHIN AN OPENSHIFT AI ENVIRONMENT?
Answer: D. MariaDB or MySQL Database
Pipeline metadata in OpenShift AI is stored in a MariaDB or MySQL database provisioned within the project. Candidates should recognize this database requirement when deploying the pipeline server rather than assuming a generic caching layer.
Q6. Which API object configuration captures Prometheus scraping endpoints, attaching comprehensive AI endpoint real-time graphing?
Answer: D. ServiceMonitor Custom Resource objects
ServiceMonitor custom resources define the metrics endpoints that Prometheus scrapes for monitoring model-serving performance. ConfigMaps and standard routes do not automatically configure Prometheus scraping targets.
Q7. Which component permits embedding specific AWS S3 credentials to a project namespace without manual re-entry per user?
Answer: D. Data Connection instance inside the OpenShift AI dashboard UI
A Data Connection securely stores S3 credentials in a Kubernetes Secret and injects them directly into your workbenches. This prevents manual re-entry, unlike generic secrets, because the dashboard handles the specific OpenShift AI environment variable mappings automatically.
Q8. Which component of OpenShift AI provides an interface to manage distributed compute tuning via Ray clusters?
Answer: D. CodeFlare Dashboard
The CodeFlare Dashboard provides the visual interface to manage, scale, and tune distributed compute resources like Ray clusters. Other options focus on model serving or pipelines, whereas CodeFlare specifically targets distributed training.
Q9. When creating a workbench, what specific configuration can be applied to inject cluster environment variables automatically?
Answer: A. ConfigMap reference
ConfigMap references are used to populate and inject environment variables into workbench pods. Using a ConfigMap is the standard Kubernetes-native method, whereas Roles and Routes handle permissions or routing instead.
Q10. What is the primary S3-compatible storage engine deployed commonly in OpenShift to supply on-premise AI data connections?
Answer: D. Red Hat OpenShift Data Foundation (ODF)
Red Hat OpenShift Data Foundation provides the S3-compatible object storage frequently used for OpenShift AI data connections. For on-premise deployments, ODF is the expected answer, whereas the other options are public cloud services or unsupported local provisioners.
Q11. Which role is capable of modifying user authorization levels within independent data science projects via the Access tabs?
Answer: D. Project Administrator role
The Project Administrator role is responsible for managing user authorizations within a Data Science Project, specifically through the Access tab. Standard cluster users lack the necessary role bindings to modify these permissions.
Q12. In OpenShift AI Pipelines, which underlying open-source engine handles the workflow execution of pipeline steps?
Answer: A. Tekton
The Dashboard server component of OpenShift AI pipelines integrates directly with Kubeflow Pipelines to orchestrate workflows. However, because OpenShift Pipelines natively relies on Tekton, exam candidates could reasonably argue Tekton provides the underlying execution.
Q13. Which custom resource creation declares an operational pre-trained machine learning specification ready to receive inference API payloads?
Answer: A. InferenceService CRD configuration spec
Creating an InferenceService custom resource declares a model ready to receive inference payloads in OpenShift AI. The controller uses this spec to provision the predictor and route traffic to the deployed runtime.
Q14. What pipeline interface module permits constructing reusable workflows interfaces visually directly within Jupyter workbench?
Answer: B. Elyra Pipeline Editor graphical pipeline canvas
The Elyra Pipeline Editor provides a graphical canvas to build reusable workflows inside a Jupyter workbench. The Kubeflow Pipelines SDK requires writing Python code instead of offering a native visual interface.
Q15. WHICH PERMISSION LEVEL MUST A USER POSSESS TO VIEW BUT NOT EDIT OTHERS' DATA SCIENCE PROJECTS?
Answer: A. Project viewer role via OpenShift AI dashboard share settings
The Project viewer role specifically grants read-only access to a project via the dashboard share settings. Cluster-reader gives broad read access across the entire cluster rather than targeted project permissions.
Q16. Which component is responsible for orchestrating serverless deployment strategies and autoscaling-to-zero for served models?
Answer: B. OpenShift Serverless based on Knative
OpenShift Serverless uses Knative Serving to manage serverless model deployments and scale pods to zero. Standard OpenShift routes maintain persistent connections and cannot scale applications based on incoming traffic.
Q17. WHICH CLUSTER OPERATOR MUST BE DEPLOYED TO ASSIST THE NVIDIA GPU OPERATOR IN DETECTING CORE ACCELERATOR HARDWARE METRIC EXPOSURES?
Answer: B. Node Feature Discovery (NFD) Operator
The Node Feature Discovery operator tags nodes with specific hardware labels like GPUs for proper scheduling. The SRIOV Network Operator handles specialized networking interfaces rather than accelerator discovery.
More Red Hat Certified Specialist in OpenShift AI (EX267) drills and other practice exams are on @CertPunch. New rounds drop every few days at certpunch.com.