1Z0-1127-25 Sample Questions & Answers
Built around the OCI Generative AI service itself, double the weight of anything else, plus large language model basics, prompt design, retrieval-augmented generation through LangChain and 23ai, and building RAG agents with knowledge bases.
Launch the full 1Z0-1127-25 simulator →Showing 8 of 17 free samples.
- Question 1Intermediate
Using OCI Generative AI RAG Agents service · Discuss options for creating knowledge bases
An administrator of the OCI Generative AI Agents service needs to delete a Knowledge Base named 'Q3-Financial-Reports'. However, every attempt to delete it from the console fails without a specific error message. What is the most probable reason for this failure and what action must be taken first?
Show answer & explanation
Correct answer: C
The OCI Generative AI Agents service maintains dependencies between resources. A Knowledge Base cannot be deleted if one or more Agents are actively using it. This is a protective measure to prevent breaking deployed applications. The correct procedure is to first delete any Agents that rely on the Knowledge Base, or edit them to use a different one, before attempting to delete the Knowledge Base itself.
- Question 2IntermediateSelect 2
Using OCI Generative AI Service · Explore OCI Generative AI security architecture
A retail company is using OCI Generative AI to create personalized product descriptions. To ensure brand consistency and prevent the model from generating unsafe content, the AI architect needs to implement security controls. Which TWO of the following OCI security mechanisms are most directly applicable to controlling access to the Generative AI service and monitoring its usage? (Select TWO)
Show answer & explanation
Correct answers: B, D
OCI Identity and Access Management (IAM) is the primary mechanism for controlling who can do what with OCI resources. Creating granular policies to allow or deny actions like creating fine-tuning jobs, deploying endpoints, or invoking models is fundamental to securing the service.
The OCI Audit service automatically records all API calls made to OCI services in a tenancy. This is crucial for monitoring who is using the Generative AI service, what actions they are performing, and when. It provides an immutable log for security analysis and compliance.
- Question 3Intermediate
Implement RAG using OCI Generative AI service · Explain RAG and RAG workflow
A data engineer is building the ingestion pipeline for a RAG system. The source consists of thousands of large PDF documents stored in an OCI Object Storage bucket. Which sequence of steps, using a framework like LangChain, correctly describes the data preparation process before the data can be stored in a vector database?
flowchart LR A[OCI Object Storage] --> B{Load Documents} B --> C{Split into Chunks} C --> D{Generate Embeddings} D --> E[(Vector Database)]Show answer & explanation
Correct answer: B
This is the correct sequence for the RAG ingestion pipeline. First, the raw documents are loaded (e.g., using a PDF loader). Second, because LLM embedding models have context limits, the loaded text is split into smaller, manageable chunks. Finally, each of these text chunks is passed to an embedding model to create its corresponding vector representation, which can then be stored.
- Question 4Intermediate
Implement RAG using OCI Generative AI service · Explain RAG and RAG workflow
A project manager is deciding between two approaches for building a question-answering system for their company's internal documentation: full fine-tuning a base model versus implementing a Retrieval-Augmented Generation (RAG) system. The documentation is updated daily with new policies and procedures. Which of the following is the STRONGEST reason to choose the RAG approach in this scenario?
Show answer & explanation
Correct answer: B
This is the core advantage of RAG for this use case. With RAG, updating the system's knowledge is as simple as updating the documents in the vector store, a process that can be easily automated. In contrast, fine-tuning embeds knowledge into the model's weights, requiring a costly and time-consuming retraining process every time the documentation changes. For frequently updated knowledge, RAG is far more practical and scalable.
- Question 5Intermediate
Using OCI Generative AI Service · Create and use model endpoints for inference
A machine learning engineer has deployed a custom fine-tuned model to a model endpoint on a Dedicated AI Cluster. During load testing, they observe that the endpoint's response time degrades significantly as the number of concurrent requests increases. The cluster's GPU utilization is high, but CPU and memory are stable. What is the most effective solution to improve performance and handle the variable load?
Show answer & explanation
Correct answer: B
The scenario describes a compute bottleneck, indicated by high GPU utilization. The solution is to scale the model endpoint horizontally by increasing the number of hosting units (instances) on the Dedicated AI Cluster. This distributes the inference workload across more GPUs, allowing the endpoint to handle more concurrent requests and reducing latency.
- Question 6Intermediate
Using OCI Generative AI RAG Agents service · Content Moderation Configuration
A developer is configuring an OCI Generative AI Agent and wants to ensure that the agent's generated responses are also checked for harmful or inappropriate content before being sent back to the user. Which specific setting within the agent's configuration must be enabled?
Show answer & explanation
Correct answer: B
The OCI Generative AI Agent configuration provides separate controls for moderating inputs and outputs. To check the agent's own generated text, the 'Moderate responses' (or similarly named) option within the Content Moderation section must be explicitly enabled. This ensures a two-way safety check.
- Question 7Advanced
Implement RAG using OCI Generative AI service · Describe similarity search and retrieve chunks from Oracle Database 23ai
A developer is building a RAG system and observes that the answers are often factually correct but lack relevance because the initial vector search returns documents that are semantically similar but contextually wrong for the specific user query. The diagram below shows a basic RAG pipeline. Where should a 'Reranker' component be added to fix this issue?
sequenceDiagram participant User participant App as Application participant VecDB as Vector DB participant LLM User->>App: Query App->>VecDB: Vector Search (Retrieval) VecDB-->>App: Top-K Chunks App->>LLM: Prompt + Chunks LLM-->>App: Answer App-->>User: Final AnswerShow answer & explanation
Correct answer: B
This is the correct placement. A reranker is a second-stage model that takes the initial list of retrieved documents (Top-K Chunks) from the vector database and re-orders them based on a more sophisticated relevance calculation with respect to the original query. This ensures that the most contextually relevant chunks are placed first before being passed to the LLM, improving the final answer's quality.
- Question 8Intermediate
Fundamentals of Large Language Models (LLMs) · Understand LLM architectures
What is the key difference between an encoder-only LLM architecture (like BERT) and a decoder-only architecture (like GPT)?
Show answer & explanation
Correct answer: B
This is the fundamental difference. Encoder-only models like BERT can see the entire input sequence at once (bidirectional context), which makes them powerful for tasks that require a deep understanding of the full context, like sentiment analysis or named entity recognition. Decoder-only models like GPT use a causal or masked self-attention mechanism, where each token can only attend to previous tokens. This autoregressive property makes them perfectly suited for generating text one token at a time.
Ready for the real thing?
The full 1Z0-1127-25 simulator has every exam-style question, timed mode, and instant scoring.