Oracle Cloud Infrastructure Generative AI Professional 1Z0-1127 (1Z0-1127) — Questions 1–75

Request throughput (requests per second) is a critical metric for monitoring the operational efficiency of a generative AI deployment on OCI because it directly measures the system's capacity to handle incoming inference requests. If throughput drops below expected levels, it indicates a bottleneck in the compute resources (e.g., GPU utilization) or the serving infrastructure, which can lead to degraded user experience and potential timeouts.

Option C is correct because the high latency and 504 timeouts with 50 concurrent users indicate that a single GPU replica is overwhelmed by the request queue. Increasing replicas to 3 distributes the load across multiple endpoints, while enabling autoscaling based on CPU utilization ensures dynamic scaling to handle traffic spikes. This directly reduces per-request latency and eliminates timeouts without violating data residency requirements.

A model must be deployed to an endpoint to be accessible via the inference API. Registration in the Model Catalog alone does not create an endpoint. Option A is correct.

The error indicates the model ID specified in the request does not exist or is misspelled.

Option A is correct because increasing the number of replicas in the dedicated cluster distributes the load across multiple model copies, improving concurrency and latency stability.

Option B is correct because OCI Monitoring service allows you to define and track custom metrics such as inference latency (e.g., p50/p99 response times) and throughput (requests per second) for your generative AI model deployed on OCI Data Science. This enables real-time performance monitoring and alerting based on thresholds you set, which is essential for production AI workloads.

Semantic chunking ensures each chunk contains a coherent idea, and chunk overlap prevents loss of context at boundaries, improving retrieval quality.

Option B is correct because manually scaling down the dedicated AI cluster when not in use directly reduces compute costs by stopping idle GPU/CPU resources. In OCI Generative AI, dedicated AI clusters incur charges for provisioned capacity, so scaling down during inactivity avoids paying for unused infrastructure.

Option A is correct because the managed serving endpoint is available in Frankfurt region, providing low latency without the need for a dedicated cluster. Dedicated AI cluster is not necessary for latency and data residency if the managed endpoint is in the same region. The playground is for testing, and Quick Actions are for pre-built models.

Option B is correct because the clarity of the prompt directly determines how well the LLM interprets the user's intent. A well-structured, unambiguous prompt reduces ambiguity and guides the model toward generating relevant and coherent responses, while a vague or poorly worded prompt often leads to off-target or nonsensical output.

Option B is correct because RAG retrieves relevant, up-to-date information from an external knowledge base at inference time, allowing the model to answer questions about recent events or proprietary data without requiring any retraining. This is a key advantage over fine-tuning, which would need a new training cycle to incorporate the same new knowledge.

Option B is correct because selecting the smallest GPU instance type that meets latency requirements directly reduces compute cost per inference without sacrificing user experience. This aligns with OCI's pay-as-you-go GPU pricing, where larger instances incur higher hourly costs. The key is to right-size the GPU based on model memory footprint and inference throughput, not to over-provision.

Reducing the k value in the nearest neighbor search directly decreases the number of vectors that must be compared during query time, which lowers latency. In approximate nearest neighbor (ANN) search, a smaller k means fewer candidates are evaluated, speeding up retrieval while still maintaining acceptable accuracy if the original k was unnecessarily high. This is the most effective tuning knob for latency in vector search systems like OCI OpenSearch with Cohere embeddings.

If k (the number of nearest neighbors to return) is set too low, the search may not find any documents within the top k, returning no results.

The error indicates that the model is not available in the specified region. OCI Gen AI models are deployed regionally, and each region supports only a specific subset of models. If the user calls an endpoint in a region where the requested model has not been deployed, the service returns an error because the model's inference endpoint does not exist in that region's routing table.

Option D is correct because reserved capacity with a fixed monthly fee provides the lowest per-request cost for high-volume batch processing. OCI Generative AI offers dedicated capacity pricing, which is ideal for predictable, large-scale workloads where you commit to a certain throughput, avoiding per-request charges that would accumulate significantly with high volume.

OCI Search with OpenSearch supports vector indexing and search natively, making it a suitable managed solution for RAG vector storage.

Option A is correct because the 'use' verb on the 'generative-ai-model' resource type grants the LegalTeam group permission to invoke the model endpoint for inference, which is the minimum privilege required for accessing a deployed fine-tuned model in OCI Generative AI Service. The 'use' permission specifically allows calling the model for text generation or analysis without granting broader management or read capabilities.

Option B is correct because a smaller fine-tuned model offers faster inference and sufficient accuracy for simple FAQs. Option A is overkill and slow, Option C may still be large, and Option D adds unnecessary complexity.

The OCI Generative AI Playground is a web-based, no-code interface that allows data scientists to interactively test prompts and adjust parameters like temperature and max tokens without writing any code. This directly matches the user's requirement for quick, code-free experimentation.

RDMA (Remote Direct Memory Access) bypasses the CPU and kernel to transfer data directly between GPU memories, drastically reducing latency and CPU overhead. In a dedicated AI cluster on OCI, high-bandwidth RDMA networking (e.g., using RoCE v2 or InfiniBand) is the most critical step to minimize inference latency because model parallelism and tensor parallelism across nodes depend on fast, low-latency interconnects. Without RDMA, even with optimized model parallelism, the network becomes the bottleneck, increasing per-token latency.

Option D is correct because query rewriting or expansion using a language model (LLM) directly addresses typos and abbreviations by generating a corrected or enriched query before embedding. This improves the semantic alignment between the user's intent and the vector search, ensuring that even noisy input retrieves relevant chunks. Techniques like spelling correction or synonym expansion at query time are far more effective than post-retrieval fixes or parameter tuning.

Option C is correct because the frequency penalty directly reduces the likelihood of the model repeating the same phrases by penalizing tokens that have already appeared in the generated text. In OCI Generative AI Service, this parameter subtracts a fixed value from the log-probability of each token each time it is generated, making repeated tokens less likely to be chosen again. This is the most direct mechanism to address repetitive output.

Using a multilingual embedding model (A) handles multiple languages in a single pipeline, and storing separate vector indices per language (C) allows optimized retrieval for each language.

A dedicated AI cluster provides isolated compute resources (GPU nodes) for inference, eliminating resource contention from other tenants or workloads. This ensures consistent low-latency responses because the model is always warm and available without queueing delays, which is critical for real-time applications.

The problem is likely chunk boundaries cutting off context. Using a sliding window with overlap ensures continuity, so relevant information is not lost at chunk edges. Increasing k adds noise; switching model is expensive; manual rephrasing is not scalable.

Creating a custom model deployment on a Dedicated AI Cluster requires the user to have the 'Use of virtual-network-family for cluster networking' permission. This permission allows the user to specify and manage the virtual network (VCN) and subnet that the cluster uses for networking. Without it, the deployment fails because the cluster cannot be attached to the required network resources.

Increasing the temperature parameter makes the model's output more random and diverse, which is ideal for creative tasks like generating marketing copy. A higher temperature (e.g., 0.7–1.0) increases the probability of sampling less likely tokens, leading to more varied and imaginative text. Setting temperature to 0 would make the output deterministic and repetitive, which is the opposite of what the user wants.

Option B is correct because deploying multiple dedicated AI clusters across different regions with a load balancer ensures high availability by distributing traffic and providing failover if one region becomes unavailable. OCI Generative AI dedicated AI clusters are provisioned per region, and a load balancer can route requests to healthy clusters, meeting the requirement for real-time document translation with cross-region redundancy.

Option A is correct because Llama 3 is one of the open-source large language models (LLMs) available through the OCI Generative AI service, alongside Cohere models. OCI Generative AI provides managed access to Llama 3 for text generation tasks, allowing users to deploy and fine-tune it within Oracle Cloud Infrastructure.

Expanding acronyms to their full forms during document preprocessing and indexing ensures that the embedding model can map the acronym to its semantic meaning, improving retrieval recall for domain-specific terms. Cohere embeddings are trained on general text, so without expansion, acronyms like 'NLP' may not match queries for 'Natural Language Processing' in vector space. This preprocessing step directly addresses the root cause of low recall for acronyms.

Using smaller chunk sizes reduces the amount of text per embedding, speeding up retrieval and subsequent processing. Increasing topK would retrieve more contexts and slow down response. Switching to a more expensive model or adding a reranker would increase latency.

Option B is correct because OCI Data Science model deployment requires an inference script (typically score.py) to define how the model processes incoming prediction requests. This script is the entry point that loads the model artifact and executes inference logic, making it an essential component for serving predictions.

Option C is correct because generative AI models deployed on OCI expose a RESTful endpoint that accepts HTTP requests (typically POST with JSON payloads) for inference. This is the standard method for programmatic access to the model, allowing integration with any HTTP client.

Increasing shards on the same node partitions the index, so each shard contains fewer vectors, making exact search faster. This reduces latency without sacrificing accuracy. ANN reduces recall, scaling adds cost and complexity, and dimension reduction can degrade embedding quality.

If the query is not converted to an embedding before search, the vector store may interpret the raw text as an invalid query or fall back to lexical search, yielding irrelevant results.

Resource principal requires that the dynamic group containing the session has a policy granting the 'use generative-ai-embeddings' permission. If the session is not in the dynamic group, the policy does not apply.

Option B is correct because fine-tuning on a dedicated AI cluster keeps data within OCI and provides accuracy for medical domain. Option A is also within OCI but less specialized; C is multi-tenant with potential isolation concerns; D uses third-party.

Option A is correct because common tokenization methods in large language models include word-based tokenization (splitting text into whole words) and subword-based tokenization (like Byte-Pair Encoding or WordPiece), which handle out-of-vocabulary words and morphological variations more effectively. Subword tokenization is widely used in models like GPT and BERT to balance vocabulary size and coverage.

Option B is correct because OCI Data Science Model Deployment provides a managed environment for hosting custom generative AI models with REST API endpoints and built-in authentication via OCI IAM. It integrates natively with OCI Monitoring and OCI Logging to track data drift and operational metrics without requiring additional infrastructure setup, minimizing operational overhead.

Dedicated AI clusters provide predictable latency and guaranteed throughput because they are single-tenant.

OCI requires specific IAM policies to allow users to use Generative AI services. A missing or incorrect policy is the typical cause of 403 errors.

When few-shot examples are too similar, the model overfits to a narrow pattern, reducing its ability to generalize to the diverse reasoning paths required by the task. This is a known limitation of in-context learning: the model treats the examples as a template rather than as diverse demonstrations. Using more diverse examples exposes the model to a wider range of reasoning patterns, improving robustness.

Option C is correct because prompt engineering allows the company to inject security constraints directly into the instruction without requiring additional training data or infrastructure. By crafting a prompt that explicitly requests secure code (e.g., 'Generate code that follows OWASP Top 10 best practices and avoids SQL injection, XSS, and buffer overflows'), the model can leverage its existing knowledge to produce safer outputs. This approach is immediate, cost-effective, and does not depend on the size or quality of the small secure code dataset.

Option A is correct because reducing the number of output tokens directly speeds up generation. Option B increases latency, Option C may not affect latency, and Option D helps throughput but not individual request latency.

Using the OCI OpenSearch document update API to replace embeddings for changed documents is efficient and targeted, ensuring immediate consistency.

Top_p (nucleus sampling) directly impacts consistency by controlling the cumulative probability threshold for token selection. A lower Top_p (e.g., 0.1) restricts the model to only the most probable tokens, reducing randomness and making outputs more deterministic across repeated prompts. This parameter, along with Temperature, is a primary lever for managing output variability in LLMs.

Option C is correct because the OCI CLI paginates results by default; using the --all flag retrieves all models. Options A and D are possible but the most common cause is pagination; B is less likely since the user is querying a specific compartment.

Dedicated AI clusters provide guaranteed compute resources (GPUs) with no multi-tenant contention, ensuring low latency and high throughput for inference workloads. Using a base model avoids the additional overhead of fine-tuning inference, which can introduce latency due to custom weight loading and optimization steps. This combination is optimal for real-time summarization of customer feedback where response time and volume are critical.

The max_tokens parameter limits the number of tokens in the generated output, directly controlling summary length.

The startup requires low latency and cost efficiency for a customer support chatbot. Instruct models like cohere.command are specifically fine-tuned to follow conversational instructions and generate concise, task-oriented responses, making them ideal for interactive chatbot applications. They balance performance and cost better than base models, which lack instruction-following capability, and embedding models, which are designed for semantic search rather than text generation.

OCI Data Science Model Deployment with GPU shapes is the best option because it provides managed, scalable, low-latency inference endpoints for LLMs. GPU shapes (e.g., VM.GPU.A10) are essential for the parallel matrix computations required by transformer-based models, and the deployment service supports auto-scaling and load balancing to maintain real-time response times.

Reinforcement Learning from Human Feedback (RLHF) directly addresses semantic correctness by fine-tuning the model using human-validated code examples. This process teaches the model to prefer outputs that are not only syntactically valid but also logically correct and aligned with developer intent, reducing semantically wrong code generation.

Fine-tuning with a balanced, curated dataset directly addresses the root cause of bias by adjusting the model's internal weights to reduce reliance on biased patterns in the original training data. This approach preserves the model's generative performance for product descriptions because it retrains only on domain-specific, unbiased examples, unlike post-processing which merely filters outputs without correcting the underlying model behavior.

Option C is correct because autoscaling with a minimum of 1 and maximum of 10 GPU instances allows the deployment to dynamically adjust capacity in response to variable traffic and spikes, ensuring cost-efficiency by scaling down during low demand and responsiveness by scaling up during peaks. OCI Data Science Model Deployment supports autoscaling policies that can be configured with GPU shapes, making it the optimal choice for a generative AI model requiring GPU acceleration.

The 429 (Too Many Requests) error indicates the API rate limit has been exceeded. In OCI Generative AI services, rate limits are enforced per tenancy and per region; the recommended action is to wait for the retry-after period and implement exponential backoff to avoid overwhelming the service.

Option B is correct because deploying a custom generative AI model via OCI Data Science Model Deployment requires packaging the model and its inference code into a standardized artifact format (e.g., pickle, ONNX). This artifact is the core input that the deployment runtime loads to serve predictions, making it an essential step in the workflow.

Option A is correct because OCI Object Storage is a supported source for training data when fine-tuning models in OCI Generative AI. The service can directly access data stored in Object Storage buckets via service-level integrations, allowing you to reference large datasets without local uploads.

Enabling reranking improves the relevance of retrieved documents by reordering them based on semantic match with the query. Without reranking, the initial vector search results may not be optimally ordered.

A dedicated AI cluster in OCI Generative AI is designed for real-time inference with low latency. When the cluster is configured for batch inference, it processes requests in batches rather than individually, which introduces queuing and processing delays that can easily exceed the 500ms target. This explains the observed 2-second latency, as batch mode prioritizes throughput over per-request response time.

Option A is correct because OCI OpenSearch with Approximate Nearest Neighbor (ANN) indexes is a best practice for vector storage and retrieval in RAG applications. ANN indexes enable efficient similarity search over high-dimensional embeddings, which is essential for retrieving relevant context from large document collections at low latency.

A is correct because the OCI Console provides a built-in chat interface that allows users to interact directly with Generative AI models without writing any code. This interface is part of the OCI Generative AI service's web-based console, enabling prompt testing and model evaluation through a graphical user interface.

Few-shot learning (option A) uses a handful of examples in the prompt to guide the model's behavior without fine-tuning. This is a form of prompt engineering specifically designed for few-shot scenarios.

Option A is correct because gradient accumulation allows the model to simulate the effect of a larger batch size without increasing memory usage. Instead of computing gradients over a single large batch, the optimizer accumulates gradients over several smaller batches before performing a weight update. This technique effectively decouples the batch size from memory consumption, enabling training on large models or high-resolution inputs that would otherwise cause an out-of-memory error.

Option B is correct because OCI APIs require request signing using an API signing key (an RSA key pair) for programmatic access. The developer must generate a key pair, upload the public key to the OCI console, and then use the private key to sign each HTTP request using the OCI Signature Version 1 algorithm (based on RFC 2104 HMAC-SHA256). This ensures authentication without transmitting secrets over the wire.