A C * A. Generation of content embeddings: Creating embeddings for all the published content is a computationally intensive process that doesn't need to happen in real-time as users interact with the system. * C. Creation of the search index: The search index, typically a vector database in a RAG system, needs to be built and updated to store the content embeddings.

Correct: Content Embedding (A) – Transform published documents into embeddings using a model (e.g., via Amazon Titan or OpenSearch ML). This can be done offline in batches, especially for static or periodically updated content like daily publications. Search Index Creation (C) – After generating embeddings, these need to be indexed (e.g., using Amazon OpenSearch or FAISS). This step can also be handled offline, as it's only needed when content updates. Wrong: B, D, and E require real-time processing for a near-real-time experience.

Its both A and C

- A — Content embeddings can be generated as a batch job whenever new content is published daily. - C — The search index (vector store) is built/updated offline from those embeddings. The remaining steps (B, D, E) must happen at query time (online) since they depend on the user's actual input: - B — User query embedding is generated live per request. - D — Retrieval happens live against the pre-built index. - E — Response generation happens live using the retrieved context.

The correct answers are A. Generation of content embeddings and C. Creation of the search index. Here's why these are the best choices: A. Generation of content embeddings: Can be done as a batch process when new content is published Computationally intensive task Not time-critical for user interaction Can be scheduled periodically Prepares content for efficient retrieval C. Creation of the search index: Can be built offline using the generated embeddings Resource-intensive process Can be updated periodically Crucial for efficient retrieval Doesn't need to happen in real-time

AC is the correct answer

Looks like there's an issue with choosing the options here as it doesn't allow to select more than one option here. But based on the requirements of this question, A and C are more relevant.

The correct options are A. Generation of content embeddings and C. Creation of the search index.

C is also the answer

A and C is correct I think

✅ Etapas que devem ser feitas offline (em lote): A. Geração de incorporações de conteúdo Isso é feito quando novos conteúdos são adicionados. Como os documentos não mudam frequentemente depois de publicados, essa etapa pode ser feita em lote. C. Criação do índice de pesquisa Após gerar as incorporações dos documentos, é necessário indexá-los para permitir busca eficiente. Isso também pode ser feito em lote, e atualizado conforme novos conteúdos são publicados.

A and C

A and C 1. Documents are published daily, the company can periodically process and embed this content in batches 2. Once embeddings are generated, they need to be indexed in a vector database