AWS AIP-C01: Prompt Management for Production GenAI Systems
A prompt used in production is part of an application's behavior, not a free-form note that a developer can change…
Explore more technical insights and certification explainers. Page 69 of 214.
Technical explainers and certification perspectives from our editorial library.
A prompt used in production is part of an application's behavior, not a free-form note that a developer can change…
Observability for a generative-AI application must explain why a business outcome happened, not only whether an API returned a successful…
Content safety for a foundation-model application must address both inappropriate output and inappropriate use of a legitimate tool. Amazon Bedrock…
A foundation model evaluation is useful only when it represents the application that will use the model. A broad language…
A generative-AI service can be inexpensive per request and still waste money if it retries repeatedly, retrieves irrelevant passages or…
Continuous delivery for a generative-AI application must treat models, prompts, retrieval indexes, tool schemas and access policies as parts of…
Retrieval augmented generation (RAG) is not a technique for teaching a foundation model confidential facts permanently. It is an application…
Agentic AI gives a model the ability to choose among bounded tools and plan several steps toward a business objective.…
Redshift Distribution Styles and Sort Keys belongs inside AWS data ingestion, transformation, storage, analytics, and governance because the topic affects…