AWS Bedrock & AI Services: free practice questions

Provisioned throughput in Amazon Bedrock allows customers to purchase a fixed number of model units, guaranteeing a consistent throughput level that prevents throttling — ideal for high-volume, predictable workloads. It does not eliminate per-token costs; customers pay for the committed capacity regardless of actual usage, which is a commitment cost, not free inference. Provisioned throughput applies to a specific model in specific regions; it does not grant cross-region access to all models. Provisioned throughput is a committed capacity purchase with a minimum term (e.g., 1 month or 6 months) and does not auto-scale — that characteristic describes on-demand pricing.

Grounding checks in Bedrock Guardrails compare the model's generated response against a provided reference source (e.g., retrieved Knowledge Base passages) and compute a grounding score. Responses that contain claims not supported by the reference content are flagged as potentially hallucinated — this is precisely what grounding checks are designed for. Detecting offensive or harmful language in user input is the role of content filters, not grounding checks. Restricting conversations to approved topics is handled by the denied topics feature. Detecting PII in model responses or user inputs is handled by sensitive information filters (PII redaction), not grounding checks.

Amazon Rekognition provides a fully managed Content Moderation API that can detect explicit, suggestive, or violent content in images and videos with no ML expertise required — this is its purpose-built use case. Amazon Bedrock with Stability AI is a generative image model used for creating images, not classifying or moderating them. Amazon SageMaker requires significant ML expertise to set up and maintain a custom image classifier. Amazon Textract is designed for extracting text and structured data from documents, not image content moderation.

The Groundedness metric in Amazon Bedrock's automatic model evaluation specifically measures whether the model's generated response is factually supported by the provided source passages — directly addressing the hallucination concern in RAG use cases. It evaluates whether claims in the response can be traced back to the retrieved context. ROUGE-L (option A) measures lexical overlap between a generated response and a reference answer; it does not assess whether the content is grounded in source documents and is better suited for summarization tasks. BERTScore (option B) measures semantic similarity but compares the response to a reference text, not to source passages — it does not assess groundedness in retrieved context. The Correctness metric (option C) using an LLM judge compares the response against a golden reference answer for factual accuracy, but this requires pre-defined reference answers and does not directly evaluate grounding against retrieved passages the way the Groundedness metric does.

Amazon Transcribe is AWS's managed automatic speech recognition (ASR) service that supports telephony audio formats (8 kHz), and its speaker diarization feature (ShowSpeakerLabels parameter) automatically identifies and labels different speakers in a recording. Amazon Comprehend processes text, not audio, and has no audio input capability. Amazon Rekognition analyzes images and videos but does not process audio recordings for speech-to-text. Amazon Bedrock with Titan multimodal supports text and images, not raw audio transcription.