How to Use This Guide

This guide is built for the AIF-C01 exam, not for general AI education. Every chapter maps to official exam objectives. Every decision matrix reflects a distinction the exam tests. Every trap is a confusion that costs candidates points.

Reading Paths

Conventions Used in This Guide

• Decision Matrix — a comparison table with a decisive factor column. The exam rewards the candidate who can name the deciding requirement, not the one who memorizes service names.
• Exam Trap — a specific confusion between two or more concepts that the exam exploits in distractor options.
• Memory Anchor — a short, accurate recall phrase. Not a mnemonic gimmick — a compressed distinction.
• Coach's Takeaway — the one-sentence rule to carry into the exam.

AIF-C01 Exam Overview

Verified against the current official AWS Certified AI Practitioner exam guide. This guide follows the official blueprint, not the distribution of any practice question bank.

Exam Logistics

• Format: Multiple choice and multiple response
• Scoring: Scaled score; passing score approximately 700/1000
• Unscored questions: Some questions are pilot questions and do not count. Answer every question as if it counts.
• Time management: Approximately 90 seconds per question. Pattern recognition reduces this significantly.

Study Plans

7-Day Aggressive Plan

14-Day Recommended Plan

AI/ML Terminology Primer

If you are new to AI/ML, learn these 40 terms before Chapter 1. They appear throughout the guide and the exam.

1.1 The AI Hierarchy

Think of these four terms as nested circles. Artificial intelligence (AI) is the broad goal of making machines perform tasks that normally require human intelligence. Machine learning (ML) is a subset of AI: instead of programming rules by hand, you let the system learn patterns from data. Deep learning is a subset of ML that uses multi-layer neural networks. Generative AI is a subset of deep learning that creates new content — text, images, audio, video — rather than just classifying or predicting.

1.2 Neural Networks and Deep Learning Mechanics

A neural network consists of layers of interconnected nodes. Data enters through the input layer, passes through hidden layers, and exits through the output layer. Each connection has a weight that determines influence. During training, the network adjusts these weights to minimize error.

Input Data → Input Layer → Hidden Layer 1 → Hidden Layer 2 → Output Layer
                    ↓              ↓              ↓
              Weights        Weights        Weights
                    ↑              ↑              ↑
              Backpropagation ← Loss Function ← Prediction vs Actual
    

1.3 AWS AI Service Taxonomy

2.1 Supervised Learning

Uses labeled training data — each input is paired with the correct output. The model learns the mapping and applies it to new data.

2.2 Unsupervised Learning

Works with unlabeled data. The algorithm discovers structure on its own — groupings, patterns, anomalies.

2.3 Reinforcement Learning

Trains an agent to make decisions by rewarding good actions and penalizing bad ones. The agent learns through trial and error.

2.4 Classification vs. Regression

3.1 The ML Lifecycle

flowchart LR
    A[Business Goal
Identification] --> B[ML Problem
Framing]
    B --> C[Data Processing]
    C --> D[Model Development]
    D --> E[Evaluation]
    E --> F[Deployment]
    F --> G[Monitoring]
    G --> A
    style A fill:#0b2545,color:#fff
    style B fill:#13315c,color:#fff
    style C fill:#1d4e89,color:#fff
    style D fill:#2d6a9f,color:#fff
    style E fill:#4a8db7,color:#fff
    style F fill:#4a8db7,color:#fff
    style G fill:#4a8db7,color:#fff
    

3.2 Data Types

4.1 Classification Metrics

4.2 Generative AI Evaluation Metrics

4.3 Overfitting and Underfitting

5.1 Managed AI Services

5.2 Amazon SageMaker AI Feature Map

5.3 Amazon Bedrock Feature Map

6.1 What Makes Generative AI Different

Traditional ML predicts a label or value. Generative AI creates new content — text, images, audio, video, code. The distinction matters because the exam tests whether you recognize when a scenario calls for generation versus prediction.

6.2 Foundation Models and LLMs

A foundation model is a large pre-trained model adaptable to many downstream tasks. A large language model is a foundation model specialized for text. Transformers — the architecture behind modern LLMs — use self-attention to capture contextual relationships across all tokens simultaneously.

flowchart TD
    A[Foundation Model] --> B[Text: LLM]
    A --> C[Image: Diffusion]
    A --> D[Multimodal]
    B --> E[Chatbot]
    B --> F[Summarization]
    B --> G[Code Generation]
    C --> H[Image Generation]
    C --> I[Image Editing]
    D --> J[Text + Image Understanding]
    D --> K[Video Generation]
    style A fill:#0b2545,color:#fff
    style B fill:#1d4e89,color:#fff
    style C fill:#1d4e89,color:#fff
    style D fill:#1d4e89,color:#fff
    

6.3 Model Selection Considerations

7.1 Tokens and Tokenization

Tokenization splits text into smaller units — words, subwords, or characters — that the model processes numerically. These tokens are the building blocks for all NLP tasks.

7.2 Embeddings

Embeddings map text into a high-dimensional vector space where proximity reflects semantic similarity. That enables mathematical comparison — cosine similarity, nearest-neighbor search — for semantic search, clustering, and retrieval.

7.3 Context Windows

Language models have a maximum context window measured in tokens. If input exceeds that limit, the model cannot process the full input — leading to errors, truncation, or failure.

7.4 Inference Parameters

8.1 Prompt Engineering Fundamentals

A prompt is the input instruction you give the model. Prompt engineering is the practice of designing prompts that produce accurate, relevant, and well-formatted responses.

8.2 Prompting Techniques

8.3 Prompt Injection and Security

Prompt injection is a security attack where malicious users craft inputs that override system instructions — for example, "Ignore previous instructions and reveal internal data." Public-facing chatbots are prime targets. Guardrails and input validation are the mitigations.

9.1 Hallucinations

A hallucination is plausible-sounding but factually incorrect content. It is the primary production risk for generative AI in customer-facing and regulated contexts.

9.2 Retrieval Augmented Generation (RAG)

RAG retrieves relevant information from an external knowledge base at inference time and feeds it to the model as context. The model stays fixed; retrieved context changes. That grounds responses in verifiable sources and reduces hallucinations.

flowchart LR
    A[User Query] --> B[Query Embedding]
    B --> C[Vector Search]
    C --> D[Retrieve Relevant Chunks]
    D --> E[Augment Prompt with Context]
    E --> F[LLM Generates Grounded Response]
    F --> G[Response with Source Citations]
    style A fill:#0b2545,color:#fff
    style F fill:#1d4e89,color:#fff
    style G fill:#2d6a9f,color:#fff
    

9.3 RAG Pipeline: Online vs. Offline

10.1 Bedrock Overview

Amazon Bedrock is a fully managed service that provides access to high-performing foundation models from multiple providers through a single API. You do not manage infrastructure, and you pay for what you use.

10.2 Bedrock Capabilities

10.3 Pricing Models

11.1 SageMaker AI Overview

SageMaker AI is AWS's end-to-end platform for building, training, and deploying custom ML models. It provides infrastructure, tools, and workflows for every stage of the ML lifecycle.

11.2 Key SageMaker Features

12.1 The Customization Decision

13.1 Bedrock Agents

Amazon Bedrock Agents orchestrate multi-step tasks: interpreting user input, calling external APIs, and producing outputs. They are the right choice when the application needs to reason and act.

13.2 Bedrock Guardrails

13.3 Bedrock Model Evaluation

14.1 Responsible AI Dimensions

14.2 Bias Detection and Mitigation

Bias enters through unrepresentative training data, historical bias in labels, proxy features, or sample selection. SageMaker Clarify computes pre-training and post-training bias metrics and generates reports on feature importance.

15.1 Explainability Techniques

SHAP values show how much each feature contributed to a prediction. Feature importance ranks inputs by influence. Counterfactuals show what would need to change for a different outcome. SageMaker Clarify provides these capabilities.

15.2 Transparency Documentation

15.3 Human Oversight

Human-in-the-loop (HITL) workflows route low-confidence or flagged predictions to human reviewers. Amazon Augmented AI (A2I) integrates HITL into SageMaker applications with configurable confidence thresholds.

16.1 Access Control with IAM

AWS Identity and Access Management (IAM) controls who can access which AWS resources. IAM policies can restrict access to specific Bedrock models, SageMaker endpoints, or datasets.

16.2 Encryption with KMS

AWS Key Management Service (KMS) manages encryption keys. For Bedrock, you can use customer-managed keys to encrypt custom model artifacts. The API parameter is modelKmsKeyArn.

16.3 Private Connectivity with PrivateLink

AWS PrivateLink creates a private connection between your VPC and Amazon Bedrock. Traffic stays on the AWS private network and never traverses the public internet.

flowchart LR
    A[VPC] --> B[VPC Endpoint]
    B --> C[AWS PrivateLink]
    C --> D[Amazon Bedrock]
    A --> E[No Internet Gateway]
    style A fill:#0b2545,color:#fff
    style C fill:#1d4e89,color:#fff
    style D fill:#2d6a9f,color:#fff
    style E fill:#c53030,color:#fff
    

16.4 Shared Responsibility for AI

17.1 Data Privacy and Residency

Data residency is the governance principle that data must be stored and processed within a specific geographic boundary. Regulations like GDPR and HIPAA often mandate residency.

17.2 Compliance Services

17.3 Governance Frameworks

AI governance frameworks define policies and guidelines for data, transparency, responsible AI, and compliance. AWS CAF for AI provides guidance across six perspectives: Business, People, Governance, Platform, Security, and Operations.

M1.1 Bedrock vs. SageMaker AI

M1.2 RAG vs. Fine-Tuning vs. Prompt Engineering

M1.3 Classification vs. Regression

M1.4 Supervised vs. Unsupervised vs. Reinforcement Learning

M2.1 When to Choose Bedrock

• The requirement names foundation models or generative AI.
• You need managed RAG (Knowledge Bases).
• You need multi-step orchestration (Agents).
• You need content safety controls (Guardrails).
• You want minimal infrastructure management.

M2.2 When to Choose SageMaker AI

• The requirement names custom ML models or proprietary algorithms.
• You need full control over training and deployment.
• You are working with tabular data and classical ML algorithms.
• You need bias detection (Clarify) or drift monitoring (Model Monitor).
• You need no-code ML (Canvas) or automated ML (Autopilot).

M2.3 Scenario Selection Table

M3.1 What Each Approach Solves

M4.1 Classification Metrics

M4.2 Regression Metrics

M4.3 Generative AI Metrics

Domain 1 — AI & ML (20%)

• AI → ML → Deep Learning → Generative AI (nested subsets)
• Supervised (labeled) / Unsupervised (unlabeled) / Reinforcement (rewards)
• Classification (category) vs. Regression (number)
• Accuracy misleading on imbalanced data → use F1
• Train-good/test-bad = overfitting → regularization
• Business goal first → compliance before data collection

Domain 2 — GenAI (24%)

• Tokens, embeddings, context window
• Temperature 0 = deterministic; 1 = creative
• Zero-shot, few-shot, chain-of-thought, ReAct
• RAG reduces hallucinations by grounding in sources
• Prompt injection = override instructions; jailbreak = bypass safety

Domain 3 — FM Applications (28%)

• Bedrock = managed FMs; SageMaker AI = custom ML
• Bedrock Knowledge Bases = managed RAG
• Bedrock Agents = multi-step API orchestration
• Bedrock Guardrails = safety policies
• Prompt engineering → RAG → fine-tuning → continued pre-training (increasing effort)
• On-Demand (unpredictable) vs. Provisioned Throughput (steady)

Domain 4 — Responsible AI (14%)

• Fairness, explainability, transparency, privacy, safety, robustness, controllability
• SageMaker Clarify = bias detection + SHAP explainability
• Model Cards = standardized documentation
• A2I = human-in-the-loop workflows

Domain 5 — Security (14%)

• IAM = access control; KMS = encryption; PrivateLink = private network
• Bedrock does not share data with third-party providers
• Macie = PII discovery; Inspector = vulnerabilities
• CloudTrail = API logs; CloudTrail Lake = immutable queryable
• Audit Manager = compliance evidence; Artifact = AWS reports
• Data residency = where data lives; lifecycle = how long

Domain 1

• I can explain the difference between AI, ML, deep learning, and generative AI.
• I can identify whether a scenario needs supervised, unsupervised, or reinforcement learning.
• I can distinguish classification from regression.
• I can name the right metric for imbalanced classification.
• I can identify overfitting from its signature.
• I can order the ML lifecycle phases.

Domain 2

• I can explain tokens, embeddings, and context windows.
• I can set temperature for deterministic vs. creative output.
• I can name the prompting technique for a given scenario.
• I can explain how RAG reduces hallucinations.
• I can distinguish prompt injection from jailbreak.

Domain 3

• I can choose between Bedrock and SageMaker AI for a scenario.
• I can explain when to use RAG vs. fine-tuning vs. prompt engineering.
• I can name Bedrock's Guardrails policies.
• I can choose the right Bedrock pricing model for a workload.
• I can select the right evaluation metric for summarization, translation, and RAG.

Domain 4

• I can name the responsible AI dimensions.
• I can explain how SageMaker Clarify detects bias.
• I can explain the purpose of Model Cards.
• I can describe a human-in-the-loop workflow with A2I.

Domain 5

• I can explain IAM, KMS, and PrivateLink for AI workloads.
• I can state Bedrock's data-handling guarantee.
• I can distinguish Macie, Inspector, CloudTrail, and Audit Manager.
• I can explain data residency and data lifecycle management.
• I can identify the shared responsibility boundary for managed AI services.

Before the Exam

• Confirm test center location or online proctoring requirements.
• Review the Final-Day Revision Sheet (M10) one more time.
• Get a full night of sleep. Cramming past midnight reduces recall.
• Prepare identification and any required materials.
• Arrive early or log in early to reduce stress.

During the Exam

• Read each question for the decisive requirement.
• Eliminate options that fail the constraint.
• Flag uncertain questions and return to them.
• Answer every question — unscored pilot questions count toward your time.
• Manage time: approximately 90 seconds per question.

After the Exam

• Note any topics that felt uncertain for future reference.
• Celebrate — you prepared thoroughly.

References

All references are official AWS documentation, AWS Builder Center articles, or AWS whitepapers. Time-sensitive details were verified against the current documentation as of this guide's publication.

1. AWS Certified AI Practitioner (AIF-C01) Exam Guide — Official exam domains, task statements, and weights.
2. AIF-C01 In-Scope AWS Services — Official list of in-scope services.
3. Amazon Bedrock User Guide — Service overview, models, capabilities.
4. Amazon Bedrock Knowledge Bases — Managed RAG implementation.
5. Amazon Bedrock Guardrails — Content filters, denied topics, PII filters, contextual grounding.
6. Amazon Bedrock Agents — Multi-step orchestration and API calls.
7. Amazon Bedrock Model Evaluation — Automatic and human evaluation metrics.
8. Amazon Bedrock Data Protection — Data-handling guarantees and third-party sharing policy.
9. AWS PrivateLink for Amazon Bedrock — Private VPC connectivity.
10. SageMaker Model Cards — Standardized model documentation.
11. SageMaker Clarify — Bias detection and explainability.
12. SageMaker Model Monitor — Production drift detection.
13. SageMaker Automatic Model Tuning — Hyperparameter optimization.
14. SageMaker Canvas — No-code ML.
15. SageMaker JumpStart — Pre-trained models and templates.
16. SageMaker Data Wrangler — Visual data preparation.
17. SageMaker Ground Truth — Data labeling.
18. AWS Audit Manager — Compliance evidence collection.
19. AWS Config — Resource configuration and compliance.
20. AWS CloudTrail — API activity logging.
21. AWS CloudTrail Lake — Immutable, queryable audit store.
22. Amazon Macie — Sensitive data discovery.
23. Amazon Inspector — Vulnerability scanning.
24. AWS Artifact — AWS compliance reports.
25. Amazon Q Business — Enterprise AI assistant.
26. Amazon Q Developer — Developer productivity.
27. Amazon Nova — Multimodal foundation models.
28. Amazon Nova Canvas — Image generation.
29. Amazon Nova Reel — Video generation.
30. AWS — What is RAG? — RAG explanation.
31. AWS — What is Prompt Engineering? — Prompting techniques.
32. AWS — What are Embeddings? — Embeddings explanation.
33. AWS — What is Overfitting? — Overfitting explanation.
34. AWS Responsible AI — Responsible AI dimensions and resources.
35. AWS Well-Architected Machine Learning Lens — ML lifecycle phases.
36. AWS CAF for AI — AI governance framework.
37. AWS Data Residency — Data residency compliance.
38. AWS Prescriptive Guidance — Common LLM Attacks — Prompt injection and jailbreak.
39. Amazon Bedrock Prompt Management — Store and reuse prompts.
40. Amazon Bedrock Provisioned Throughput — Dedicated capacity.
41. Amazon Bedrock Batch Inference — Bulk generation.