AIF-C01 cheat sheet

Exam at a glance

Domain weightings

How much of the exam each domain covers. Spend your study time in proportion — the heavier the domain, the more questions you'll see.

Key terms

Foundation Model

A foundation model is a large machine-learning model pre-trained on broad data that can be adapted to many tasks such as text generation, summarization, and Q&A. AIF-C01 covers selecting and applying foundation models through services like Amazon Bedrock.

Amazon Bedrock

Amazon Bedrock is a fully managed service that provides access to foundation models from multiple providers through a single API, with features for RAG, agents, and guardrails. It is the core generative-AI building block on the AIF-C01 exam.

Prompt Engineering

Prompt engineering is the practice of designing inputs to guide a foundation model toward the desired output using instructions, examples, and context. AIF-C01 covers how prompt design affects quality and how to reduce undesirable output.

Retrieval Augmented Generation

Retrieval Augmented Generation (RAG) is a technique that supplies a foundation model with relevant retrieved documents at inference time so its answers are grounded in your data. AIF-C01 contrasts RAG with fine-tuning as ways to customize model behavior.

Fine-tuning

Fine-tuning is the process of further training a foundation model on labeled task-specific data to specialize it. AIF-C01 covers when fine-tuning is appropriate versus using RAG or prompt engineering, considering cost and effort.

Amazon SageMaker

Amazon SageMaker is AWS's managed platform for building, training, and deploying machine-learning models. AIF-C01 references it as the primary AWS service across the ML lifecycle, including tools like SageMaker Clarify for bias detection.

Machine Learning

Machine learning is a branch of AI in which systems learn patterns from data rather than being explicitly programmed. AIF-C01 covers supervised, unsupervised, and reinforcement learning and their appropriate use cases.

Inference

Inference is the process of using a trained model to make predictions or generate output on new input data. AIF-C01 covers inference concepts including latency, cost, and real-time versus batch inference.

Hallucination

A hallucination is a confident but incorrect or fabricated output from a generative model. AIF-C01 covers hallucination as a limitation of generative AI and mitigations such as RAG, guardrails, and human review.

Responsible AI

Responsible AI is the practice of building AI systems that are fair, transparent, robust, and safe. AIF-C01 covers its dimensions — bias, explainability, transparency — and AWS tools like SageMaker Clarify and Guardrails for Amazon Bedrock.

SageMaker Clarify

Amazon SageMaker Clarify is a tool that helps detect bias in data and models and explains model predictions. AIF-C01 covers it as a key service for the responsible-AI and fairness objectives.

Embedding

An embedding is a numerical vector representation of text, images, or other data that captures semantic meaning for similarity search and retrieval. AIF-C01 covers embeddings as the mechanism behind RAG and vector search.

Token

A token is the basic unit of text a foundation model processes, roughly a word fragment. AIF-C01 covers tokens because they drive model cost and context limits in generative-AI applications.