SaveMyCert
Log in
5 of 5 free questions left today·for 10 a day
AIF-C01 · Domain 1

Fundamentals of AI and ML practice questions

Fundamentals of AI and ML is worth 20% of the AIF-C01 exam — the 3rd-heaviest of the 5 domains. Core AI and ML concepts and terminology, practical use cases, and the AI/ML development lifecycle. 6 fully worked examples are further down this page, answers included.

Exam weight
20%
the 3rd-heaviest of the 5 domains
Questions
84
across 3 topics
Free, no account
5/day
a free account allows 10 a day
Explanations
Every option
right and wrong

Build a practice session

5 free questions left today.

Domains

How many?

5 free questions left today.

Mode

Ready when you are

10 fresh questions drawn across 1 of 5 domains, in Learn mode.

Review what you missed

Every question you answer incorrectly, and every question you flag while practising, is saved here automatically. Finish a session and you can come back to re-drill just those.

6 sample Fundamentals of AI and ML questions, fully explained

Questions from the AIF-C01 bank mapped to domain 1, with the answer key and the reasoning behind every option. None of them repeat the examples on the main AIF-C01 practice page.

Question 1Fundamentals of AI and ML

A company builds two assistants on the same foundation model. The first answers each question it is given. The second breaks a request such as 'arrange a team offsite' into steps, calls a calendar API and a booking API, remembers earlier results, and checks each outcome before continuing. Which term best describes the second assistant?

Choose one.

  • a
    An agentic AI system Correct

    It uses a generative model as its reasoning engine and adds planning, memory, and tool calls to take multi-step actions toward a goal, which is what defines agentic AI.

  • b
    A reinforcement learning system

    Reinforcement learning is a way of training a model from rewards in an environment; the assistant here is not being trained, it is planning and acting at run time.

  • c
    A multimodal generative AI system

    Multimodal means a model handles more than one input or output type, such as text and images; it says nothing about planning or calling tools.

  • d
    A retrieval-augmented generation system

    RAG retrieves documents to ground a single answer; it does not plan steps or take actions through external APIs.

The concept

Agentic AI is generative AI that plans, keeps memory, and uses tools to act toward a goal over multiple steps.

Why that’s the answer

Both assistants use the same foundation model, so what separates the second one is behavior: it decomposes a goal, calls external tools, carries results forward, and checks progress. That plan-act-observe loop is agentic AI. Reinforcement learning is tempting because the assistant 'checks outcomes', but RL describes how a model is trained, not how a deployed assistant acts. RAG adds retrieved knowledge to one response and multimodal describes input/output types, neither of which involves taking actions.

How to reason it out
  1. Note that the underlying model is the same, so the difference is in how the system behaves.
  2. Spot the signals: multi-step planning, tool calls, memory, and checking results.
  3. Match a generative model that plans and acts with tools to agentic AI.
  4. Rule out terms that describe training methods, grounding, or data types.

Exam tip: GenAI answers; agentic AI plans and acts with tools toward a goal.

AI vs ML vs Deep Learning: Core AIF-C01 Concepts and Terminology — the lesson that teaches this.

Question 2Fundamentals of AI and ML

An algorithm processes example data and produces a trained artifact that stores the learned patterns and can make predictions. What is that trained artifact called?

Choose one.

  • a
    An algorithm

    The algorithm is the method or procedure used to learn, not the trained result it produces.

  • b
    A dataset

    A dataset is the input data used for training, not the trained artifact that comes out.

  • c
    A model Correct

    A model is the trained result produced by running an algorithm over data, and it is what makes predictions.

  • d
    A feature

    A feature is an individual input variable, not the overall trained artifact.

The concept

An algorithm is the learning method; a model is the trained result it produces.

Why that’s the answer

The trained artifact that captures learned patterns and makes predictions is the model. The algorithm is the tempting near-twin, but it is the learning procedure (the recipe), not the result it produces (the cake). A dataset is the raw input and a feature is one input variable, so neither is the trained artifact.

How to reason it out
  1. Identify the algorithm as the procedure that learns from data.
  2. Identify the model as the output of training that procedure.
  3. Match the trained, prediction-making artifact to the term model.

Exam tip: The algorithm is the recipe; the model is the finished result.

AI vs ML vs Deep Learning: Core AIF-C01 Concepts and Terminology — the lesson that teaches this.

Question 3Fundamentals of AI and ML

A finished fraud-detection model receives a new, previously unseen transaction and returns a prediction. Which phase of the machine learning workflow is this?

Choose one.

  • a
    Training

    Training is the earlier phase where the model learns patterns from example data, before it is used on live input.

  • b
    Inferencing Correct

    Inferencing is the phase where a finished model is given new input and produces a prediction.

  • c
    Data labeling

    Data labeling attaches correct answers to examples during data preparation, not when the model scores new input.

  • d
    Feature engineering

    Feature engineering shapes raw data into useful inputs during preparation, not the moment a model makes a prediction.

The concept

Training builds a model; inferencing uses the finished model on new data.

Why that’s the answer

Scoring a new, unseen transaction with an already-trained model is inferencing. Training is the earlier phase in which the model learned from historical examples. Data labeling and feature engineering are data-preparation steps that happen before training, so neither describes the moment a finished model returns a prediction.

How to reason it out
  1. Check whether the model is still learning or already finished.
  2. A finished model producing a prediction on new input is inferencing.
  3. Reserve training for the earlier learning phase.

Exam tip: Making predictions on new input with a finished model is inferencing.

AI vs ML vs Deep Learning: Core AIF-C01 Concepts and Terminology — the lesson that teaches this.

Question 4Fundamentals of AI and ML

A team has thousands of emails, each already marked as spam or not-spam, and wants a model to predict the label for future emails. Which type of machine learning does this describe?

Choose one.

  • a
    Unsupervised learning

    Unsupervised learning uses unlabeled data, but here every example already carries a correct label.

  • b
    Reinforcement learning

    Reinforcement learning uses rewards from an environment rather than a fixed set of labeled examples.

  • c
    Supervised learning Correct

    The training data is labeled with known answers (spam or not-spam), which is the defining trait of supervised learning.

  • d
    Self-supervised learning

    Self-supervised learning creates its own training signal from unlabeled data (for example, predicting hidden words); here humans have already supplied a label for every email, which is supervised learning.

The concept

Supervised learning trains on labeled data where each example includes the correct answer.

Why that’s the answer

Every email already carries the correct answer (spam or not-spam), and the goal is to predict that label for new emails, which is supervised learning. Self-supervised learning is the near-twin, but it derives its own training signal from unlabeled data rather than using human-provided labels. Unsupervised learning has no labels at all, and reinforcement learning learns from rewards in an environment.

How to reason it out
  1. Check whether each training example already has the correct answer attached.
  2. Labeled examples used to predict a label for new data means supervised learning.
  3. Rule out learning types that use no labels, self-generated signals, or rewards.

Exam tip: Labeled examples with known answers signal supervised learning.

AI vs ML vs Deep Learning: Core AIF-C01 Concepts and Terminology — the lesson that teaches this.

Question 5Fundamentals of AI and ML

A security team wants a model to discover unusual patterns in login activity on its own, with no labeled examples of normal or suspicious behavior. Which type of machine learning is this?

Choose one.

  • a
    Unsupervised learning Correct

    The data has no labels, so the model must find structure and anomalies on its own, which is unsupervised learning.

  • b
    Supervised learning

    Supervised learning needs labeled examples, but here no normal or suspicious labels are provided.

  • c
    Reinforcement learning

    Reinforcement learning relies on rewards from interacting with an environment, not on finding structure in a static unlabeled dataset.

  • d
    Transfer learning

    Transfer learning reuses a pre-trained model on a related task; it does not describe learning from unlabeled data with no answers.

The concept

Unsupervised learning finds structure in unlabeled data with no correct answers provided.

Why that’s the answer

With no labels available and the goal of discovering unusual patterns, the model must find structure by itself, which is unsupervised learning. Supervised learning would need examples already labeled normal or suspicious, reinforcement learning would need an environment that hands out rewards, and transfer learning describes reusing a pre-trained model rather than a way of learning from unlabeled data.

How to reason it out
  1. Confirm the data has no labels or known answers.
  2. Match learning that discovers structure on its own to unsupervised learning.
  3. Eliminate options that require labels, rewards, or a pre-trained source model.

Exam tip: No labels plus finding hidden structure means unsupervised learning.

AI vs ML vs Deep Learning: Core AIF-C01 Concepts and Terminology — the lesson that teaches this.

Question 6Fundamentals of AI and ML

A logistics company wants software that improves its warehouse picking routes by trying different routes in a simulated warehouse and receiving a score after each attempt. No dataset of correct routes exists. Which type of machine learning fits this approach?

Choose one.

  • a
    Supervised learning

    Supervised learning needs a dataset of labeled examples (here, known correct routes), which the stem says does not exist.

  • b
    Unsupervised learning

    Unsupervised learning finds groupings or structure in a static unlabeled dataset; it does not improve by acting and being scored.

  • c
    Self-supervised learning

    Self-supervised learning generates its own labels from a large unlabeled dataset; it does not learn by taking actions in an environment and receiving feedback.

  • d
    Reinforcement learning Correct

    An agent acting in an environment and learning from a score (reward) after each attempt, with no labeled answers, is reinforcement learning.

The concept

Reinforcement learning trains an agent by trial and error, using rewards from an environment instead of labeled data.

Why that’s the answer

The software acts (chooses a route), the simulated environment responds with a score, and the software adjusts to earn better scores. That reward-driven trial and error with no dataset of correct answers is reinforcement learning. Supervised learning would need labeled correct routes, while unsupervised and self-supervised learning both work from a fixed dataset rather than from feedback on actions.

How to reason it out
  1. Check whether a labeled dataset exists; here it does not.
  2. Notice the system takes actions and receives a score for each one.
  3. Match learning from rewards in an environment to reinforcement learning.

Exam tip: Actions plus rewards, with no labeled answers, means reinforcement learning.

AI vs ML vs Deep Learning: Core AIF-C01 Concepts and Terminology — the lesson that teaches this.

What AIF-C01 domain 1 tests, topic by topic

The official exam guide breaks Fundamentals of AI and ML into 3 topics. The question bank follows the same split, so a weak topic shows up as a cluster of misses you can go back and read.

Published AIF-C01 practice questions per topic in Fundamentals of AI and ML
TopicWhat it coversQuestions
Explain basic AI concepts and terminologiesExam guide task 1.1 (AIF-C01). Defining basic AI terms (AI, ML, deep learning, neural networks, computer vision, natural language processing [NLP], model, algorithm, training and inferencing, bias, fairness, fit, large language model [LLM], generative AI [GenAI], agentic AI); similarities and differences between AI, ML, GenAI, deep learning, and agentic AI; types of inferencing (batch, real-time, asynchronous, serverless); types of data in AI models (labeled and unlabeled, tabular, time-series, image, text, structured and unstructured); types of AI/ML learning (supervised, unsupervised, reinforcement learning methods).28
Identify practical use cases for AIExam guide task 1.2 (AIF-C01). Recognizing where AI/ML provides value (assisting human decision making, solution scalability, automation); when AI/ML solutions are not appropriate (cost-benefit analyses, when a specific outcome is needed instead of a prediction); selecting AI/ML techniques for use cases (regression, classification, clustering); real-world AI applications (computer vision, NLP, speech recognition, recommendation systems, fraud detection, forecasting, knowledge bases, agentic AI); capabilities of AWS managed AI/ML services (Amazon SageMaker AI, Amazon Transcribe, Amazon Translate, Amazon Comprehend, Amazon Lex, Amazon Polly); when traditional ML models or foundation models (FMs) fit a use case (regulatory concerns, explainability requirements, operational constraints).28
Describe the AI/ML development lifecycleExam guide task 1.3 (AIF-C01). Components of an AI/ML pipeline; sources of FM models (open source pre-trained models, training custom models); methods to use a model in production (managed API service, self-hosted API); relevant AWS services for each pipeline stage (Amazon Bedrock, Amazon Q, Amazon Quick, Kiro, SageMaker AI); fundamental MLOps concepts (experimentation, repeatable processes, scalable systems, managing technical debt, production readiness, model monitoring, model re-training); model performance metrics (accuracy, precision, recall, F1 score) and business metrics (cost per user, development costs, customer feedback, return on investment [ROI]).28
Total84

Revise Fundamentals of AI and ML before you drill it

Other AIF-C01 domains

Fundamentals of AI and ML: your questions

Fundamentals of AI and ML is domain 1 of the AIF-C01 exam guide and carries 20% of the scored content — the 3rd-heaviest of the 5 domains. On a 65-question paper that works out to roughly 13 questions, though AWS does not publish an exact per-domain count and individual exam forms vary.

Source

The domain weight and topic list on this page come from the official AIF-C01 exam guide.