Home / AIF-C01 / Objective 3.1

Domain 3: Applications of Foundation Models · 28% of the exam · Full exam outline

3.1 Describe design considerations for applications that use foundation models (FMs)

What this objective covers · 7% of the exam

Identify selection criteria to choose FMs (for example, cost, modality, latency, multi-lingual, model size, model complexity, customization, input/output length, prompt caching).; Describe the effect of inference parameters on model responses (for example, temperature, input/output length).; Define Retrieval Augmented Generation (RAG) and describe its business applications (for example, Amazon Bedrock Knowledge Bases).; Identify AWS services that help store embeddings within vector databases (for example, Amazon OpenSearch Service, Amazon Aurora, Amazon Neptune, Amazon RDS for PostgreSQL).; Explain the cost tradeoffs of various approaches to FM customization (for example, pre-training, fine-tuning, in-context learning, RAG, model distillation).; Define the role of AI agents and describe AI agents' business applications.

Practice questions

Question 1easy

A developer wants the model to stop generating as soon as it writes the marker "###" at the end of each record. Which inference setting provides this behavior?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

Temperature makes output more or less deterministic. It has no notion of a marker at which to stop.

Penalties discourage repeated or frequent tokens in a response. They do not end the output when a particular marker appears.

Top P controls how much of the probability distribution the model samples from for each token. It changes word choice, not where generation ends.

A stop sequence is a string that tells the model to stop generating further tokens once it produces that string, which is exactly the behavior wanted for the record marker.

Question 2easy

A water utility wants its customer chatbot to answer from tariff documents it publishes every quarter. The FM never saw these documents during training, and the company does not want to retrain the model. Which approach matches this need?

Included in Master: master every domain.See plans