AIF-C01 — AWS Certified AI Practitioner Cheat Sheet

Last revised: September 16, 2026

Cheat sheet: AWS Certified AI Practitioner (AIF-C01) reference for AI/ML concepts, AWS services, generative AI, security, governance, and exam decision points.

Use the tables for a quick pre-exam check. Expand a topic’s notes for explanations, examples, and additional distinctions.

Scope and study context

The AIF-C01 exam is foundational. Expect questions that test whether you can:

Explain core AI, ML, and generative AI concepts.
Match business use cases to appropriate AWS services.
Recognize responsible AI, security, privacy, and governance concerns.
Understand high-level model lifecycle, data preparation, evaluation, and monitoring.
Choose between managed AI services, Amazon Bedrock, and Amazon SageMaker based on scenario clues.

This page is IT Mastery review support and is not affiliated with AWS.

Exam focus at a glance

This independent Cheat Sheet supports preparation for the AWS Certified AI Practitioner (AIF-C01) exam from AWS. Use it to quickly connect AI/ML concepts, AWS service choices, responsible AI, and security/governance decisions.

What to recognize quickly	Exam-prep cue
AI vs ML vs deep learning vs generative AI	AI is the broad field; ML learns from data; deep learning uses neural networks; generative AI creates new content using foundation models.
Managed AI service vs custom ML model	Use managed AI services for common tasks; use Amazon SageMaker when you need custom training, tuning, deployment, or MLOps.
Amazon Bedrock vs Amazon SageMaker	Bedrock is for building with foundation models and generative AI APIs; SageMaker is for building, training, tuning, deploying, and monitoring ML models.
Prompt engineering vs RAG vs fine-tuning	Prompting changes instructions; RAG adds retrieved knowledge; fine-tuning changes model behavior for a task or style.
Responsible AI risks	Bias, hallucination, toxicity, data leakage, explainability gaps, and misuse.
Security responsibilities	AWS secures the cloud infrastructure; customers secure identities, data, prompts, access, model use, monitoring, and compliance configuration.
Evaluation metrics	Match metric to problem type: precision/recall/F1 for classification, RMSE/MAE for regression, grounding/toxicity/human review for generative AI.

Core AI, ML, and generative AI vocabulary

Term	Compact meaning	High-yield distinction
Artificial intelligence	Systems that perform tasks associated with human intelligence	Broad umbrella: includes ML, rules, optimization, NLP, vision, robotics
Machine learning	Models learn patterns from data	Not explicitly programmed rules for every case
Deep learning	ML using multi-layer neural networks	Common for language, images, speech, complex patterns
Foundation model	Large model trained on broad data and adaptable to many tasks	Base for generative AI, chat, summarization, embeddings, code, image tasks
Large language model	Foundation model focused on language	Generates, summarizes, classifies, extracts, reasons over text
Multimodal model	Works with more than one modality	Text plus images, audio, video, or documents
Token	Unit of text processed by a model	Cost, latency, and context capacity are affected by token usage
Embedding	Numeric vector representing semantic meaning	Used for similarity search, clustering, recommendations, and RAG
Prompt	Input instructions and context sent to a model	Better prompts can improve output without retraining
Inference	Using a trained model to make predictions or generate output	Different from training or fine-tuning
RAG	Retrieval-Augmented Generation	Retrieves trusted context before generation; useful for private/current facts
Fine-tuning	Additional training on labeled or curated examples	Best for task behavior, style, or format; not ideal for constantly changing facts
Hallucination	Plausible but incorrect model output	Mitigate with grounding, retrieval, validation, guardrails, and human review
Bias	Systematic unfairness or skew in data/model output	Can come from data, labels, sampling, features, or feedback loops
Feature	Input variable used by a model	Feature quality strongly affects ML performance
Label	Known target value for supervised learning	Example: fraud/not fraud, price, category
Ground truth	Trusted correct answer or label	Needed for training and evaluation
Training set	Data used to fit model parameters	Should not include validation/test leakage
Validation set	Data used to tune model choices	Helps select features, hyperparameters, thresholds
Test set	Held-out data for final evaluation	Should represent real future data
Overfitting	Performs well on training data but poorly on new data	Often too complex, too little data, or leakage
Underfitting	Model too simple to capture patterns	Poor training and validation performance
Drift	Data or relationship changes over time	Requires monitoring and possible retraining
MLOps	Operational practices for ML lifecycle	Includes versioning, deployment, monitoring, rollback, governance

AWS AI service selection matrix

Need	Primary AWS service to consider	Choose when	Common trap
Build generative AI app with foundation models	Amazon Bedrock	You need managed access to FMs, chat, summarization, RAG, agents, embeddings, or guardrails	Confusing Bedrock with full custom ML training infrastructure
Build, train, tune, deploy, and monitor custom ML	Amazon SageMaker	You need notebooks, training jobs, model registry, endpoints, batch inference, monitoring, or MLOps	Choosing Bedrock when the scenario requires custom training pipelines
No-code/low-code ML for business users	Amazon SageMaker Canvas	Users need predictions without writing code	Assuming every ML scenario requires writing training code
Enterprise assistant over business data	Amazon Q Business	You need a managed workplace assistant connected to enterprise content	Using Bedrock directly when the requirement is a ready business assistant
Developer coding assistant	Amazon Q Developer	You need code suggestions, explanations, or development assistance	Confusing developer productivity assistant with a general ML platform
Text sentiment, entities, key phrases, language detection	Amazon Comprehend	You need managed NLP extraction/classification	Using a foundation model for simple standard NLP when a specialized service fits
Speech to text	Amazon Transcribe	Convert audio into text	Confusing with Amazon Polly, which creates speech from text
Text to speech	Amazon Polly	Generate natural-sounding speech from text	Confusing with Transcribe
Translate text	Amazon Translate	Machine translation across languages	Using custom LLM prompts when a managed translation service is enough
Chatbot/contact center conversational interface	Amazon Lex	Build voice/text conversational bots	Confusing with Bedrock chat; Lex manages intents, slots, and bot flow
Enterprise search	Amazon Kendra	Search across enterprise documents with relevance and connectors	Confusing with generic vector search or log/search analytics
Search, analytics, vector similarity	Amazon OpenSearch Service	Need search indexes, log analytics, vector search, similarity retrieval	Choosing Kendra when you need lower-level search/vector control
Extract text, tables, forms from documents	Amazon Textract	OCR plus structured document extraction	Confusing with Rekognition image/video labels
Image/video labels, faces, moderation	Amazon Rekognition	Computer vision for images/video	Using Textract for object/face detection
Personalized recommendations	Amazon Personalize	Recommendations without building the full ML stack	Assuming recommendation systems always require SageMaker
Human data labeling	Amazon SageMaker Ground Truth	Need labeled datasets with workflows and human review	Treating unlabeled raw data as ready for supervised learning
Data catalog and ETL	AWS Glue	Prepare/catalog data for analytics or ML	Confusing data preparation with model training
Object storage for datasets/artifacts	Amazon S3	Store raw data, training data, model artifacts, logs	Not automatically a catalog, access governance, or model registry
Data lake permissions	AWS Lake Formation	Govern data lake access and permissions	Assuming S3 bucket policies alone cover all lake governance needs
Data discovery/governance collaboration	Amazon DataZone	Share, catalog, discover, and govern data across teams	Confusing with model monitoring or AI guardrails

Notes and examples

Managed AI services vs Amazon Bedrock vs Amazon SageMaker

Choice	Use when…	Candidate mistake to avoid
Managed AWS AI service	The task is standard and specific: OCR, speech, translation, sentiment, image labels, chatbot intents	Over-engineering with custom ML
Amazon Bedrock	You need foundation models, generative AI, embeddings, RAG, agents, or guardrails through managed APIs	Treating Bedrock as a traditional custom model training platform
Amazon SageMaker	You need to prepare data, train, tune, deploy, monitor, or manage custom ML models	Choosing SageMaker when a simpler managed AI API satisfies the use case
Amazon SageMaker Canvas	Business users need no-code or low-code predictions	Assuming every SageMaker scenario requires data scientists writing code
Amazon Q Business	Organization wants a generative AI assistant connected to company data and business apps	Confusing enterprise assistant use cases with Amazon Lex
Amazon Q Developer	Developers want coding, AWS guidance, or software development assistance	Confusing with business-user knowledge assistant use cases

AWS AI and ML services at a glance

Service	Primary use	Fast exam cue
Amazon Bedrock	Build generative AI applications with foundation models	LLMs, embeddings, RAG, agents, guardrails
Amazon SageMaker	Build, train, tune, deploy, and monitor ML models	Full ML lifecycle
Amazon SageMaker Canvas	No-code ML for business analysts	Predictions without writing code
Amazon SageMaker Ground Truth	Data labeling workflows	Human labeling, annotation
Amazon SageMaker Clarify	Bias detection and model explainability	Fairness, explainability
Amazon SageMaker Model Monitor	Monitor deployed model quality and drift	Production ML monitoring
Amazon Textract	Extract printed/handwritten text, forms, and tables from documents	OCR plus document structure
Amazon Comprehend	NLP for text insights	Sentiment, entities, key phrases, language
Amazon Transcribe	Speech to text	Audio becomes text
Amazon Polly	Text to speech	Text becomes audio
Amazon Translate	Language translation	Translate between languages
Amazon Lex	Conversational interfaces using voice/text	Intents, slots, chatbot flow
Amazon Rekognition	Image and video analysis	Objects, scenes, faces, moderation labels
Amazon Personalize	Personalized recommendations	User-item recommendations
Amazon Kendra	Intelligent enterprise search	Search across internal documents
Amazon OpenSearch Service	Search, analytics, and vector search patterns	Semantic search, vector retrieval
Amazon Q Business	Generative AI assistant for enterprise knowledge	Business assistant over company data
Amazon Q Developer	Generative AI assistant for developers	Code, AWS development help
AWS Glue	ETL and Data Catalog	Prepare/catalog data
Amazon S3	Object storage for data lakes, datasets, artifacts	Durable storage foundation
AWS Lake Formation	Data lake governance	Permissions and governance for data lakes
Amazon Athena	Query S3 data using SQL	Serverless interactive query
Amazon Redshift	Data warehouse analytics	Large-scale structured analytics
Amazon QuickSight	Business intelligence dashboards	Visualize and share insights
Amazon Macie	Discover sensitive data in S3	PII/sensitive data detection
AWS IAM	Identity and access control	Least privilege
AWS KMS	Encryption key management	Protect data at rest
AWS CloudTrail	API activity audit logs	Who did what, when
Amazon CloudWatch	Metrics, logs, alarms	Operational monitoring

Fast decision path: managed AI, generative AI, or custom ML

    flowchart TD
	    A[AI requirement] --> B{Common prebuilt task?}
	    B -->|Speech, OCR, vision, translation, NLP| C[Use specialized AWS AI service]
	    B -->|No| D{Need generated language, chat, summarization, embeddings, agents?}
	    D -->|Yes| E{Needs private or current knowledge?}
	    E -->|Yes| F[Amazon Bedrock + RAG / Knowledge Bases]
	    E -->|No| G[Amazon Bedrock prompt, FM, or agent]
	    D -->|No| H{Need custom predictive model?}
	    H -->|Business user/no code| I[SageMaker Canvas]
	    H -->|Code, training, tuning, MLOps| J[Amazon SageMaker]
	    H -->|No| K[Re-check problem; may be analytics, automation, or search]

Amazon Bedrock and generative AI reference

Bedrock concept	What it does	Exam decision cue
Foundation models	Pretrained models available through a managed service	Choose based on task, modality, quality, latency, cost, context needs, and customization support
Model inference	Send prompt/input and receive output	Core operation for chat, generation, summarization, classification, extraction
Embeddings model	Converts content into vectors	Required for semantic search and many RAG designs
Knowledge Bases for Amazon Bedrock	Managed RAG workflow over enterprise data sources/vector stores	Use when the scenario says ground responses in private documents
Agents for Amazon Bedrock	Orchestrates tasks and can call tools/APIs	Use when the model must take actions or complete multi-step workflows
Guardrails for Amazon Bedrock	Applies safety, topic, content, and policy controls	Use to reduce unsafe, off-topic, or noncompliant outputs; not a replacement for IAM
Model customization	Adapts supported models using your data	Use when prompt/RAG are insufficient for behavior, domain language, or output style
Model evaluation	Compares models and outputs against criteria	Use before selecting or changing models; include quality, safety, latency, and cost
Prompt management	Version and manage prompts	Useful when prompts are application assets requiring repeatability and change control
Inference parameters	Control output style and length	Temperature/top-p affect variability; max tokens affects length/cost/latency

Notes and examples

Foundation model selection checklist

Factor	Ask this
Task fit	Is the model strong for chat, summarization, code, extraction, image, or embeddings?
Modality	Does it support required inputs and outputs: text, image, audio, document, code?
Quality	Does it meet accuracy, reasoning, grounding, and formatting requirements?
Latency	Can it respond within user or workflow expectations?
Cost efficiency	Can a smaller or more specialized model meet the requirement?
Context handling	Can the model handle the needed prompt plus retrieved context?
Safety	Are guardrails, moderation, and evaluation adequate for the use case?
Customization	Does the model support the needed fine-tuning or customization method?
Governance	Can access, logging, data handling, and compliance expectations be met?

Prompt engineering quick reference

Pattern or parameter	Use it for	Watch for
Clear task instruction	Tell the model exactly what to do	Vague prompts lead to broad or inconsistent output
Role/context	“Act as a support analyst…”	Role alone does not guarantee correctness
Delimiters	Separate instructions, source text, examples, and output schema	Reduces prompt confusion and injection risk
Few-shot examples	Show desired input/output behavior	Bad examples teach bad behavior
Output schema	Request JSON, table, bullet list, or strict fields	Validate output; models can still produce malformed structures
Grounding context	Provide authoritative content to answer from	Ask model to say when context is insufficient
Step decomposition	Break complex tasks into smaller steps	Do not expose sensitive reasoning if not needed
Low temperature	More consistent, deterministic-style outputs	May reduce creativity
Higher temperature	Brainstorming or varied creative outputs	Can increase inconsistency or hallucination
Top-p	Controls sampling diversity	Do not randomly tune without evaluation
Max tokens	Limits response length	Too low can truncate; too high can increase cost/latency
Stop sequences	End generation at desired markers	Misconfigured stops can cut off valid output
Negative instructions	Tell the model what to avoid	Pair with positive instructions and validation

Notes and examples

Prompt safety traps

Trap	Mitigation
Putting secrets in prompts	Do not send credentials, private keys, or unnecessary sensitive data
Trusting user-provided instructions inside documents	Treat retrieved or uploaded content as untrusted data, not system instructions
Assuming prompt wording is a security control	Use IAM, authorization, validation, logging, and guardrails
Relying on “always answer correctly” wording	Use grounding, citations, evaluation, and human review
Overloading prompts with irrelevant context	Retrieve only relevant chunks and keep prompts focused

RAG: Retrieval-Augmented Generation

RAG is high yield for AIF-C01 because it connects generative AI with enterprise data, embeddings, vector search, and hallucination reduction.

RAG step	What happens	AWS-oriented cue
Ingest	Collect documents/data	Amazon S3 and enterprise data sources are common starting points
Prepare	Clean, split, and chunk content	Chunk size affects retrieval quality and prompt cost
Embed	Convert chunks to vectors	Use an embeddings model, often through Amazon Bedrock
Store/index	Save vectors and metadata	Vector-capable stores or managed knowledge base integrations
Retrieve	Find relevant chunks for a query	Similarity search plus filters improves relevance
Augment prompt	Add retrieved context to the prompt	Tell model to answer only from provided context when needed
Generate	Foundation model produces response	Include citations or references when required
Evaluate	Measure grounding and answer quality	Test with real questions and known answers
Monitor	Track failures, drift, latency, cost	Update indexes when source content changes

Notes and examples

RAG vs fine-tuning vs prompt-only

Technique	Best when	Not best when
Prompt-only	Task is simple; instructions and examples fit in context	Need large private knowledge base or persistent behavior change
RAG	Need current, proprietary, or citeable facts	Need model to learn a new style deeply or change internal behavior
Fine-tuning	Need consistent task behavior, tone, format, or domain adaptation	Facts change frequently; retrieval would be easier
Continued pretraining	Need broader domain language adaptation for supported models	Small task-specific changes or quick iteration
Full custom ML	Need model architecture/control/training pipeline beyond FM application patterns	A managed FM app already satisfies the use case

Machine learning problem types

Problem type	Goal	Example	Common metric family
Binary classification	Predict one of two classes	Fraud/not fraud	Accuracy, precision, recall, F1, ROC-AUC
Multiclass classification	Predict one of many classes	Ticket category	Accuracy, macro/micro F1
Regression	Predict numeric value	House price, demand	MAE, MSE, RMSE, R-squared
Clustering	Group similar items without labels	Customer segments	Silhouette, business validation
Anomaly detection	Find unusual observations	Suspicious transactions	Precision/recall, false positive rate
Forecasting	Predict future time-based values	Inventory demand	MAE, MAPE, RMSE
Recommendation	Suggest relevant items	Product recommendations	Click-through, ranking metrics, conversion
Natural language processing	Understand/process text	Sentiment, entities	Accuracy/F1 or task-specific metrics
Computer vision	Analyze images/video	Defect detection	Precision/recall, IoU for detection
Reinforcement learning	Learn actions through rewards	Optimization/control	Reward over time, policy performance

Learning approaches

Approach	Uses labeled data?	Output	Exam cue
Supervised learning	Yes	Predict label or value	Classification and regression
Unsupervised learning	No	Discover structure	Clustering, dimensionality reduction, anomaly discovery
Semi-supervised learning	Some	Uses small labeled plus large unlabeled data	Useful when labels are expensive
Reinforcement learning	Uses rewards	Policy/action strategy	Agent learns through interaction
Self-supervised learning	Labels derived from data itself	Representations/foundation models	Common in large-scale pretraining
Transfer learning	Starts from pretrained model	Adapts to new task	Reduces data/training needs
Active learning	Model selects examples needing labels	Efficient labeling	Useful when human labels are costly

Amazon SageMaker lifecycle reference

Lifecycle phase	SageMaker/AWS capability	What to know for the exam
Data storage	Amazon S3	Common storage for datasets, features, artifacts, logs
Data prep	AWS Glue, SageMaker Data Wrangler	Clean, transform, and prepare data
Labeling	SageMaker Ground Truth	Human labeling workflows for supervised ML
Exploration	SageMaker Studio / notebooks	Development environment for data scientists
No-code ML	SageMaker Canvas	Business users can build predictions without coding
Automated ML	SageMaker Autopilot	Automatically builds and compares candidate models
Training	SageMaker training jobs	Managed infrastructure for model training
Tuning	Hyperparameter tuning	Searches for better model settings
Experiment tracking	SageMaker Experiments	Tracks runs, parameters, metrics
Model packaging	Model artifacts and containers	Needed before deployment
Registry/governance	SageMaker Model Registry	Version, approve, and manage models
Real-time inference	SageMaker endpoints	Low-latency online predictions
Batch inference	SageMaker Batch Transform	Offline predictions over batches
Intermittent/serverless inference	SageMaker Serverless Inference	Variable or unpredictable traffic patterns
Long-running/large async requests	SageMaker Asynchronous Inference	Decoupled request/response workloads
Monitoring	SageMaker Model Monitor	Detect data quality and drift issues
Bias/explainability	SageMaker Clarify	Analyze bias and feature attribution/explainability
Pipelines	SageMaker Pipelines	Automate ML workflows and CI/CD-style steps

Evaluation metrics

Confusion matrix terms

Term	Meaning
True positive	Model predicted positive, and actual is positive
False positive	Model predicted positive, but actual is negative
True negative	Model predicted negative, and actual is negative
False negative	Model predicted negative, but actual is positive

Notes and examples

Classification metric selection

Metric	Use when	Trap
Accuracy	Classes are balanced and error costs are similar	Misleading with imbalanced classes
Precision	False positives are costly	High precision can miss many real positives
Recall	False negatives are costly	High recall can increase false positives
F1 score	Need balance between precision and recall	Hides whether precision or recall matters more
ROC-AUC	Compare ranking quality across thresholds	Can be less intuitive for severe class imbalance
Confusion matrix	Need error breakdown by class	Not a single optimization metric

\[ \begin{aligned} \text{Precision} &= \frac{TP}{TP + FP} \\ \text{Recall} &= \frac{TP}{TP + FN} \\ \text{F1} &= 2 \cdot \frac{\text{Precision} \cdot \text{Recall}}{\text{Precision} + \text{Recall}} \end{aligned} \]

Regression metric selection

Metric	Use when	Trap
MAE	Need average absolute error in original units	Treats all errors linearly
MSE	Want to penalize larger errors more	Squared units are harder to interpret
RMSE	Want original units while penalizing large errors	Sensitive to outliers
R-squared	Need explained variance view	Can be misleading alone

\[ \begin{aligned} \text{MAE} &= \frac{1}{n}\sum_{i=1}^{n} |y_i - \hat{y}_i| \\ \text{MSE} &= \frac{1}{n}\sum_{i=1}^{n} (y_i - \hat{y}_i)^2 \\ \text{RMSE} &= \sqrt{\frac{1}{n}\sum_{i=1}^{n} (y_i - \hat{y}_i)^2} \end{aligned} \]

Generative AI evaluation

Evaluation area	What to check
Relevance	Does the response answer the user’s question?
Groundedness	Is the answer supported by provided context or trusted sources?
Factuality	Are statements correct?
Hallucination rate	How often unsupported or fabricated claims appear
Toxicity/safety	Whether output violates safety or content policies
Bias/fairness	Whether outputs treat groups unfairly
Robustness	Whether output remains acceptable across prompt variations
Format adherence	Whether JSON, tables, or required fields are valid
Latency	Whether response time meets requirements
Cost efficiency	Whether token/model choices are appropriate
Human preference	Human review or pairwise comparison where automated metrics are insufficient

Foundation model concepts

Concept	What to remember
Prompt	Input instructions and context given to a generative model
System instruction	Higher-level behavior or constraints for the model
Context window	Amount of input/output text the model can consider at once
Temperature	Controls randomness; lower is more predictable, higher is more varied
Top-p	Controls sampling from probable tokens
Max tokens	Limits output length and affects cost/latency
Stop sequence	Text pattern that tells generation to stop
Embeddings	Vector representations used for semantic similarity and retrieval
Hallucination	Plausible but incorrect or unsupported output
Grounding	Tying model output to trusted context or source data
RAG	Retrieval-Augmented Generation: retrieve relevant content, then generate an answer using it
Fine-tuning	Adapting a model’s behavior using task-specific examples
Agent	System that uses a model to reason over tasks and call tools/APIs
Guardrail	Control to reduce unsafe, unwanted, or noncompliant outputs

RAG vs fine-tuning vs prompt engineering

Need	Best first approach	Why
Improve instructions, format, tone, or constraints	Prompt engineering	Fastest and lowest operational change
Answer using current or private company knowledge	RAG	Adds external context without retraining the model
Reduce hallucinations by grounding in approved documents	RAG plus evaluation and guardrails	The model can cite or use retrieved sources
Teach a repeated task style or domain-specific output pattern	Fine-tuning	Changes behavior based on examples
Add new factual knowledge that changes often	RAG	Easier to update documents than retrain
Enforce safety boundaries	Guardrails plus prompt controls	Do not rely on prompt wording alone
Connect model to actions or APIs	Agent architecture	Model can plan and invoke tools under controls

Typical RAG flow

Store trusted documents in a searchable knowledge source.
Convert document chunks into embeddings.
Store embeddings in a vector-capable store.
Convert the user query into an embedding.
Retrieve the most relevant chunks.
Add retrieved context to the prompt.
Generate a grounded response.
Apply guardrails, logging, evaluation, and human review where needed.

Common RAG traps:

Poor chunking can retrieve irrelevant or incomplete context.
Stale source documents produce stale answers.
Retrieval does not guarantee correctness; evaluate generated answers.
RAG helps with knowledge grounding but does not automatically solve authorization. Users should only retrieve data they are allowed to access.
Prompt injection can occur when retrieved content contains malicious instructions. Guardrails and input/output controls matter.

Classification metrics

Know what each metric favors. You usually do not need heavy math, but you should understand the tradeoff.

Metric	Plain meaning	Use when…	Trap
Accuracy	Overall percent correct	Classes are balanced and errors have similar cost	Misleading for imbalanced data
Precision	Of predicted positives, how many were correct	False positives are costly	High precision may miss true positives
Recall	Of actual positives, how many were found	False negatives are costly	High recall may create many false positives
F1 score	Balance of precision and recall	Need one combined metric	Hides which error type matters more
ROC-AUC	Ranking/separation quality across thresholds	Comparing binary classifiers	Does not directly pick the operating threshold
Confusion matrix	Counts true/false positives/negatives	Understanding error types	Must interpret positive class correctly

Useful formulas:

\[ \text{Accuracy} = \frac{\text{correct predictions}}{\text{all predictions}} \]\[ \text{Precision} = \frac{\text{true positives}}{\text{true positives} + \text{false positives}} \]\[ \text{Recall} = \frac{\text{true positives}}{\text{true positives} + \text{false negatives}} \]\[ \text{F1} = 2 \times \frac{\text{precision} \times \text{recall}}{\text{precision} + \text{recall}} \]

Regression and generative AI evaluation

Evaluation area	Metric or method	What it tells you
Regression error	MAE	Average absolute error; easier to explain
Regression error	RMSE	Penalizes large errors more strongly
Regression fit	R-squared	Amount of variance explained
Generative AI quality	Human evaluation	Whether output is useful, accurate, and appropriate
Generative AI grounding	Factuality/groundedness checks	Whether response is supported by source context
Generative AI safety	Toxicity, harmful content, policy checks	Whether output violates safety requirements
Generative AI relevance	Relevance scoring	Whether output answers the user’s question
Operations	Latency, throughput, error rate	Whether the solution performs in production
Cost	Cost per request, token usage, infrastructure cost	Whether the solution is economically viable

Responsible AI reference

Risk	What it looks like	Mitigation options
Bias	Unequal performance across groups	Representative data, bias testing, SageMaker Clarify, human review
Hallucination	Confident false answer	RAG, citations, grounding checks, refusal rules, human approval
Toxic output	Harmful, unsafe, abusive content	Amazon Bedrock Guardrails, moderation, testing, policy controls
Data leakage	Sensitive data appears in prompts/logs/outputs	Data classification, redaction, IAM, encryption, logging controls
Prompt injection	User or document tries to override instructions	Separate trusted instructions from untrusted content; validate tool calls
Overreliance	Users accept AI output without judgment	Human-in-the-loop for high-impact decisions
Lack of explainability	Cannot justify prediction	Interpretable models, feature attribution, documentation
Model drift	Performance degrades over time	Monitoring, retraining, alerts, evaluation baselines
Data poisoning	Training/retrieval data is manipulated	Source validation, access controls, lineage, review workflows
Inappropriate use	Model used outside intended context	Usage policies, access control, monitoring, user education

Notes and examples

Responsible AI principles to connect to scenarios

Principle	Practical exam interpretation
Fairness	Test for disparate impact and biased outcomes
Explainability	Provide understandable reasons or feature influence where needed
Privacy	Minimize, protect, and control sensitive data
Safety	Prevent harmful, toxic, or unsafe outputs/actions
Transparency	Document model purpose, limitations, data, and intended use
Robustness	Validate behavior under varied inputs and edge cases
Accountability	Assign ownership, approvals, monitoring, and escalation paths
Human oversight	Keep humans in control for high-risk or ambiguous decisions

Responsible AI review

AIF-C01 candidates should recognize responsible AI as a lifecycle concern, not a single feature.

Theme	Meaning	Practical controls
Fairness	Avoid unjustified performance gaps or harmful bias	Representative data, bias checks, SageMaker Clarify, human review
Explainability	Understand why a model produced an output	Feature attribution, interpretable models, documentation
Transparency	Communicate AI use, limitations, and confidence appropriately	User notices, model documentation, clear escalation paths
Privacy	Protect personal and sensitive data	Data minimization, masking, encryption, access control
Security	Protect systems, models, data, and prompts	IAM, KMS, network controls, logging, secure APIs
Safety	Reduce harmful, toxic, or inappropriate outputs	Guardrails, content filters, testing, human oversight
Robustness	Maintain quality under realistic input variation	Evaluation, adversarial testing, monitoring
Governance	Manage approvals, accountability, and auditability	Policies, versioning, logs, risk review, ownership
Controllability	Keep humans and systems in control of AI behavior	Constraints, approval workflows, rollback options

Common responsible AI mistakes:

Treating fairness as only a data science issue. It also involves product design, monitoring, and governance.
Assuming a model is objective because it is mathematical.
Using sensitive data without a clear purpose or access controls.
Deploying generative AI without testing for hallucinations, unsafe output, and prompt injection.
Failing to document known limitations.
Ignoring human review for high-impact or ambiguous decisions.

Security and governance for AI workloads on AWS

Control area	AWS services/features	Exam-prep distinction
Identity and authorization	AWS Identity and Access Management, IAM roles, policies	Least privilege controls who can invoke models, access data, deploy endpoints
Central identity	AWS IAM Identity Center	Workforce access management; not a model evaluation tool
Encryption at rest	AWS Key Management Service, service encryption features	Protect datasets, logs, model artifacts, vector stores
Encryption in transit	TLS/service endpoints	Protect data moving between clients and services
Network isolation	VPC design, security groups, private connectivity where supported	Reduce public exposure for data and inference paths
Secrets	AWS Secrets Manager, AWS Systems Manager Parameter Store	Do not hardcode API keys or database credentials
Logging	AWS CloudTrail, Amazon CloudWatch Logs	CloudTrail records API activity; CloudWatch collects metrics/logs/alarms
Monitoring	Amazon CloudWatch, SageMaker Model Monitor	System health plus model/data quality monitoring
Data discovery	Amazon Macie	Find and help protect sensitive data in Amazon S3
Threat detection	Amazon GuardDuty	Detect suspicious activity; not an AI quality tool
Configuration governance	AWS Config	Track resource configuration and compliance rules
Organization guardrails	AWS Organizations, service control policies	Restrict actions across accounts
Data lake governance	AWS Lake Formation	Fine-grained data lake permissions
AI output controls	Amazon Bedrock Guardrails	Safety/content controls; not a substitute for IAM, encryption, or monitoring

Notes and examples

Shared responsibility reminder

AWS is responsible for	Customer is responsible for
Security of AWS infrastructure	Data classification and lawful/appropriate use
Managed service infrastructure operations	IAM policies, roles, access boundaries
Physical facilities and core cloud platform	Prompt content, uploaded data, and retrieval sources
Availability of AWS service controls	Configuration of logging, encryption, monitoring, and guardrails
Service security features	Evaluating outputs, handling bias, hallucination, and misuse risks

Data foundations for AI on AWS

Data need	AWS service	Why it matters for AI
Durable object storage	Amazon S3	Common location for raw, curated, training, and inference data
Catalog and ETL	AWS Glue	Makes data discoverable and transformable
Query data in S3	Amazon Athena	Serverless SQL queries for data exploration
Data warehouse	Amazon Redshift	Analytics at scale; can feed ML and business intelligence
Streaming data	Amazon Kinesis, Amazon MSK	Real-time event ingestion for analytics/ML pipelines
Operational relational data	Amazon RDS, Amazon Aurora	Structured application data
Key-value/noSQL data	Amazon DynamoDB	Low-latency application data and metadata
Search/vector retrieval	Amazon OpenSearch Service	Search, log analytics, vector similarity
Data governance	AWS Lake Formation, Amazon DataZone	Permissions, discovery, sharing, governance
Sensitive data discovery	Amazon Macie	Helps identify sensitive data before using it in AI workflows

Deployment and inference patterns

Pattern	Choose when	AWS cue
Real-time inference	User or application needs immediate response	SageMaker endpoint or Bedrock model invocation
Batch inference	Large set of predictions can run offline	SageMaker Batch Transform or batch application workflow
Asynchronous inference	Requests are large or processing takes longer	Decoupled request handling with async pattern
Serverless inference	Traffic is intermittent or unpredictable	Avoid managing always-on capacity where supported
Edge inference	Need local/low-latency processing near devices	Consider edge-capable deployment patterns
API-backed GenAI app	Application calls model through API	Common with Amazon Bedrock, AWS Lambda, Amazon API Gateway
Human-in-the-loop	Decisions need review or labeling	SageMaker Ground Truth or workflow approval patterns
MLOps pipeline	Repeatable build/test/deploy/monitor	SageMaker Pipelines, Model Registry, CI/CD integration

Cost and performance levers

Lever	Why it matters
Choose the right model size	Smaller/specialized models may be faster and cheaper if quality is sufficient
Reduce unnecessary tokens	Shorter prompts and outputs can reduce latency and cost
Use retrieval carefully	More context can improve grounding but increases prompt size
Cache repeated outputs where appropriate	Reduces repeated inference for identical or stable requests
Batch offline work	Often more efficient than real-time calls for noninteractive workloads
Monitor utilization	Avoid idle always-on resources when usage is intermittent
Tune thresholds	Classification threshold changes precision/recall tradeoff without retraining
Evaluate before scaling	Do not scale a poor prompt/model/pipeline before measuring quality
Automate cleanup	Remove unused notebooks, endpoints, artifacts, and logs according to policy
Track business metric	Accuracy alone may not prove business value

High-yield AWS distinctions

Distinction	Choose this when	Not this when
Amazon Bedrock vs SageMaker	Bedrock for managed foundation model apps	SageMaker for full custom ML lifecycle
Amazon Q Business vs Bedrock	Q Business for ready enterprise assistant	Bedrock for custom GenAI application/platform control
Amazon Q Developer vs Bedrock	Q Developer for coding assistance	Bedrock for building your own app
Textract vs Rekognition	Textract extracts document text/forms/tables	Rekognition detects labels, objects, faces, video content
Transcribe vs Polly	Transcribe converts speech to text	Polly converts text to speech
Translate vs Comprehend	Translate changes language	Comprehend analyzes text meaning/entities/sentiment
Kendra vs OpenSearch	Kendra for managed enterprise search relevance	OpenSearch for search infrastructure, logs, vector search control
RAG vs fine-tuning	RAG for current/private factual knowledge	Fine-tuning for behavior, style, or task adaptation
Guardrails vs IAM	Guardrails influence model content/safety	IAM controls who can access what
CloudWatch vs CloudTrail	CloudWatch monitors metrics/logs/alarms	CloudTrail records API activity
Model Monitor vs Clarify	Model Monitor detects drift/data quality issues	Clarify helps with bias and explainability
Accuracy vs F1	Accuracy for balanced classes	F1 for precision/recall balance, especially imbalance
Precision vs recall	Precision reduces false positives	Recall reduces false negatives

Common AIF-C01 scenario cues

If the question says…	Think…
“Summarize internal policy documents accurately”	Amazon Bedrock with RAG/Knowledge Bases; evaluate groundedness
“Reduce hallucinations using company documents”	RAG, citations, retrieval quality, guardrails, human review
“Build a chatbot with intents and slots”	Amazon Lex
“Convert call recordings to text”	Amazon Transcribe
“Generate spoken audio from text”	Amazon Polly
“Extract fields from invoices or forms”	Amazon Textract
“Detect objects or unsafe image content”	Amazon Rekognition
“Analyze sentiment and entities in text”	Amazon Comprehend
“Business user wants no-code predictions”	SageMaker Canvas
“Data scientist needs training jobs and model endpoints”	Amazon SageMaker
“Need human labelers for training data”	SageMaker Ground Truth
“Need to compare model versions and approve deployment”	SageMaker Model Registry
“Need to detect drift after deployment”	SageMaker Model Monitor
“Need to analyze bias or feature attribution”	SageMaker Clarify
“Need a workplace assistant over enterprise data”	Amazon Q Business
“Need code suggestions in IDE”	Amazon Q Developer
“Need log/metric alarms”	Amazon CloudWatch
“Need record of who called what API”	AWS CloudTrail
“Need find sensitive data in S3”	Amazon Macie
“Need restrict accounts from using certain actions”	AWS Organizations service control policies

Data quality and model quality traps

Trap	Why it matters
Training/test leakage	Inflates evaluation; model may fail in production
Imbalanced classes	Accuracy can look high while minority class performance is poor
Nonrepresentative data	Model performs poorly for real users or future conditions
Label noise	Supervised model learns incorrect patterns
Missing values	Can bias results or break pipelines if untreated
Outliers	Can distort regression and distance-based models
Correlation mistaken for causation	ML predictions do not automatically prove cause
Optimizing only technical metrics	Business outcome, risk, fairness, and cost may be unacceptable
Ignoring threshold selection	Same model can behave very differently at different classification thresholds
No monitoring after deployment	Drift, data changes, and degradation go unnoticed

Quick pre-exam checklist

Can you distinguish Amazon Bedrock, Amazon SageMaker, Amazon Q, and specialized AWS AI services?
Can you choose between prompt engineering, RAG, fine-tuning, and custom ML?
Can you match metrics to classification, regression, and generative AI evaluation?
Can you explain precision vs recall and when false positives or false negatives matter?
Can you identify responsible AI controls for bias, hallucination, toxicity, privacy, and explainability?
Can you separate IAM/security controls from model safety guardrails?
Can you identify the role of CloudWatch, CloudTrail, KMS, Macie, Lake Formation, and AWS Organizations in AI governance?
Can you recognize when a scenario needs human review, data labeling, model monitoring, or retraining?

Exam mindset: how to choose the best answer

Many AIF-C01 questions are scenario-based. Do not answer only from memorized service names. First identify the task, then the level of customization, then the operational responsibility.

If the scenario says…	Think first…	Common trap
“Extract text, forms, or tables from scanned documents”	Amazon Textract	Choosing Amazon Rekognition just because an image is involved
“Analyze sentiment, entities, key phrases, or language in text”	Amazon Comprehend	Choosing Amazon Bedrock when a managed NLP service is enough
“Convert speech to text”	Amazon Transcribe	Confusing with Amazon Polly
“Convert text to lifelike speech”	Amazon Polly	Confusing with Amazon Transcribe
“Translate text between languages”	Amazon Translate	Confusing translation with summarization
“Build a conversational bot with intents and slots”	Amazon Lex	Choosing a general LLM when the exam emphasizes intent-based bot design
“Use foundation models through an API without managing infrastructure”	Amazon Bedrock	Choosing Amazon SageMaker by default
“Train, tune, build, deploy, or monitor custom ML models”	Amazon SageMaker	Choosing Bedrock when the task is custom ML lifecycle work
“No-code or low-code ML predictions for business users”	Amazon SageMaker Canvas	Choosing full SageMaker Studio-style development
“Enterprise generative AI assistant over business data”	Amazon Q Business or Bedrock with retrieval	Treating all chatbots as Amazon Lex
“Developer coding assistant”	Amazon Q Developer	Confusing with Amazon Q Business
“Search internal enterprise content”	Amazon Kendra or retrieval architecture	Confusing keyword search, semantic search, and generative answering

High-yield concepts to know cold

AI, ML, deep learning, and generative AI

Concept	Quick definition	What AIF-C01 may test
Artificial intelligence	Broad field of systems performing tasks associated with human intelligence	AI is the umbrella term
Machine learning	Systems learn patterns from data rather than being explicitly programmed for every rule	Training data, features, labels, evaluation
Deep learning	ML using neural networks with many layers	Often used for images, speech, NLP, and foundation models
Generative AI	AI that creates new content such as text, images, code, or summaries	Prompts, tokens, foundation models, hallucinations, responsible use
Foundation model	Large model trained on broad data and adaptable to many tasks	Often accessed through Amazon Bedrock
Large language model	Foundation model focused on language tasks	Summarization, Q&A, generation, reasoning-like responses
Embedding	Numeric vector representation of meaning	Search, recommendations, similarity, RAG
Token	Unit of text processed by a model	Cost, latency, context window, output length
Inference	Using a trained model to make predictions or generate output	Production use, latency, throughput, cost
Training	Learning model parameters from data	Requires data, compute, evaluation, iteration

Notes and examples

Supervised, unsupervised, and reinforcement learning

Learning type	Uses	Examples	Watch for
Supervised learning	Learn from labeled examples	Classification, regression	Needs labeled training data
Unsupervised learning	Find structure without labels	Clustering, anomaly patterns, dimensionality reduction	No “correct label” in training data
Reinforcement learning	Learn actions using rewards or penalties	Optimization, game-like decision environments, agent policies	Not the default answer for ordinary prediction
Semi-supervised learning	Mix of small labeled data plus larger unlabeled data	Reducing labeling effort	Useful when labels are expensive
Self-supervised learning	Model learns from data structure itself	Many foundation model pretraining approaches	Often foundational for generative AI

Classification, regression, clustering, and anomaly detection

Task	Output	Example	Best metric clue
Classification	Category/class	Fraud vs not fraud, image label, sentiment class	Accuracy, precision, recall, F1, ROC-AUC
Regression	Numeric value	Price, demand, wait time	MAE, RMSE, R-squared
Clustering	Groups	Customer segments	Silhouette score, business usefulness
Anomaly detection	Unusual events	Unusual transactions, abnormal sensor readings	False positives vs missed anomalies
Recommendation	Ranked items	Products, media, content	Click-through, conversion, ranking metrics

ML lifecycle quick review

AIF-C01 usually tests lifecycle understanding at a conceptual level: what happens before, during, and after model development.

    flowchart LR
	  A[Define business problem] --> B[Collect and govern data]
	  B --> C[Prepare, clean, label, and split data]
	  C --> D[Train or select model]
	  D --> E[Evaluate against metrics]
	  E --> F[Deploy for inference]
	  F --> G[Monitor quality, drift, latency, and cost]
	  G --> H[Retrain, tune, or improve]
	  H --> C

Notes and examples

Stage	Know this	Common trap
Define problem	Convert business goal into ML task and success metric	Starting with a model before defining success
Collect data	Data must be relevant, permitted, representative, and high quality	Assuming more data always fixes poor data quality
Label data	Supervised learning needs correct labels	Ignoring label noise and inconsistent annotation
Prepare data	Clean, normalize, transform, handle missing values, remove duplicates	Accidentally introducing data leakage
Split data	Use training, validation, and test data appropriately	Evaluating on the same data used to train
Train/select model	Choose model based on task, data, cost, latency, and explainability	Picking the largest model by default
Evaluate	Use metrics aligned with business risk	Relying on accuracy for imbalanced data
Deploy	Make model available for inference	Ignoring latency, scale, and security
Monitor	Watch for drift, degraded quality, bias, errors, and cost	Treating deployment as the finish line
Improve	Tune, retrain, add data, change prompts, or redesign	Changing the model without measuring impact

Data concepts that commonly appear

Data types and storage patterns

Data concept	Meaning	AWS-related clue
Structured data	Rows and columns with schema	Databases, warehouses, SQL analytics
Semi-structured data	Flexible structure such as JSON, logs, XML	Data lakes, Glue, Athena
Unstructured data	Text, images, audio, video, documents	S3, Textract, Comprehend, Rekognition, Transcribe
Data lake	Central storage for raw and processed data	Amazon S3 plus governance/catalog tools
Data warehouse	Optimized analytics on structured data	Amazon Redshift
Data catalog	Metadata about data assets	AWS Glue Data Catalog
Feature	Input variable used by a model	Customer age, text embedding, transaction amount
Label	Correct answer used in supervised learning	Fraud/not fraud, category, price
Feature engineering	Transforming data into useful model inputs	Scaling, encoding, extracting features
Data leakage	Training uses information unavailable at prediction time	Inflated test results, poor real-world performance
Data drift	Input data distribution changes over time	Monitoring and retraining needed
Concept drift	Relationship between inputs and target changes	Model may become stale even if pipeline works

Notes and examples

Data quality and bias checks

High-yield review points:

Representative data matters. If training data excludes important populations, conditions, products, geographies, or use cases, predictions may be biased or unreliable.
Labels must be accurate. Bad labels create bad supervised models.
Missing values need deliberate handling. Dropping records may bias the dataset; imputing values may introduce assumptions.
Outliers are not always errors. In fraud or anomaly detection, unusual points may be the signal.
PII and sensitive data require controls. Use data minimization, access control, encryption, masking/redaction where appropriate, and auditability.
Training and test sets must remain separate. If the model “sees” test data during training or tuning, evaluation is not trustworthy.

Security, privacy, and governance decision rules

Core AWS controls

Requirement	AWS control to consider	Exam cue
Restrict who can call a service or access data	AWS IAM	Least privilege, roles, policies
Encrypt data at rest	AWS KMS with service encryption features	Key management, encryption
Protect data in transit	TLS/HTTPS	Secure communication
Audit API activity	AWS CloudTrail	Who called which API
Monitor metrics and logs	Amazon CloudWatch	Alarms, logs, dashboards
Detect sensitive data in S3	Amazon Macie	PII discovery
Govern data lake access	AWS Lake Formation	Data lake permissions
Avoid hardcoded secrets	AWS Secrets Manager	Secure secret storage
Private connectivity to supported services	VPC endpoints / AWS PrivateLink patterns	Avoid public internet paths where required
Control S3 access	Bucket policies, IAM, encryption, block public access	Protect datasets and artifacts

Notes and examples

AI-specific security concerns

Concern	Why it matters	Mitigation direction
Prompt injection	Malicious input tries to override instructions	Input validation, guardrails, isolation, retrieval controls
Data leakage	Sensitive data appears in prompts, logs, or outputs	Data minimization, redaction, access control
Unauthorized retrieval	RAG returns documents a user should not see	Enforce permissions before retrieval and generation
Hallucinated authority	Model fabricates policies, citations, or facts	Grounding, citations, human review, evaluation
Model drift	Production behavior degrades over time	Monitoring, retraining, rollback
Over-permissioned agents	Agent can perform actions beyond user intent	Least privilege, scoped tools, approvals
Unsafe output	Harmful, biased, or noncompliant content	Guardrails, filters, testing, escalation

Cost, performance, and operational tradeoffs

AIF-C01 questions may include practical constraints such as budget, latency, scale, and maintainability.

Decision factor	What to remember
Model size	Larger models may improve quality but often increase cost and latency
Token volume	More input/output tokens usually increase cost and response time
Context length	Longer context can help but may add cost and noise
Prompt quality	Better prompts can improve results without changing models
RAG retrieval quality	Good retrieval can reduce hallucinations and improve relevance
Batch vs real time	Batch processing can be cheaper or simpler when immediate response is not needed
Managed services	Reduce operational burden for common AI tasks
Monitoring	Needed for errors, latency, drift, quality, and cost
Human review	Adds cost but may be necessary for high-risk or low-confidence outputs
Right-sizing	Match solution complexity to business value and risk

High-yield scenario patterns

Scenario clue	Likely answer direction
“Business users want predictions without coding”	Amazon SageMaker Canvas
“Data scientists need to build, train, and deploy a custom model”	Amazon SageMaker
“Use multiple foundation models through a managed service”	Amazon Bedrock
“Add enterprise documents to a generative AI Q&A workflow”	RAG, Knowledge Bases-style architecture, or Amazon Q Business depending on wording
“Prevent harmful generative AI responses”	Guardrails, content filtering, evaluation, human review
“Find sensitive data in S3 before using it for ML”	Amazon Macie
“Catalog and prepare data for analytics or ML”	AWS Glue and AWS Glue Data Catalog
“Query data directly in S3 with SQL”	Amazon Athena
“Central data lake governance”	AWS Lake Formation
“Analyze call recordings by converting audio to text”	Amazon Transcribe, then text analysis if needed
“Extract fields from invoices or forms”	Amazon Textract
“Detect objects or moderation labels in images”	Amazon Rekognition
“Identify sentiment and entities in customer reviews”	Amazon Comprehend
“Create natural-sounding audio from text”	Amazon Polly
“Build a bot that collects required fields from users”	Amazon Lex
“Translate support content into another language”	Amazon Translate
“Personalized product recommendations”	Amazon Personalize
“Audit who accessed AI resources”	AWS CloudTrail
“Encrypt data used by AI workloads”	AWS KMS and service-level encryption settings

Common candidate mistakes

Choosing the most advanced service instead of the most appropriate service. If a managed AI service directly solves the use case, it is often the best foundational answer.
Confusing Amazon Bedrock and Amazon SageMaker. Bedrock is the first thought for managed foundation model and generative AI application patterns. SageMaker is the first thought for custom ML lifecycle work.
Using fine-tuning when RAG is the better fit. If the problem is “answer from current company documents,” think retrieval and grounding before fine-tuning.
Using accuracy for imbalanced classification. A fraud model that predicts “not fraud” almost every time may have high accuracy and still be useless. Think precision, recall, F1, and business cost of errors.
Ignoring data leakage. If future information appears in training data, evaluation results may look excellent but fail in production.
Treating deployment as the end. Real systems require monitoring for drift, quality, latency, errors, security, and cost.
Assuming generative AI output is always correct. LLMs can hallucinate. Use grounding, evaluation, guardrails, citations, and human review where appropriate.
Forgetting authorization in RAG. Retrieval must respect user permissions. A model should not expose documents just because they exist in the vector store.
Confusing speech, text, and language services. Transcribe is speech-to-text. Polly is text-to-speech. Translate changes language. Comprehend analyzes text.
Overlooking responsible AI. Fairness, privacy, security, explainability, safety, robustness, transparency, and governance are all testable themes.

Fast final review checklist

Before starting topic drills or a mock exam, make sure you can answer these without hesitation:

Can you explain AI vs ML vs deep learning vs generative AI?
Can you distinguish supervised, unsupervised, and reinforcement learning?
Can you identify classification, regression, clustering, recommendation, and anomaly detection scenarios?
Can you explain features, labels, training, validation, testing, and inference?
Can you identify overfitting, underfitting, data leakage, drift, and bias?
Can you choose between managed AI services, Amazon Bedrock, and Amazon SageMaker?
Can you explain embeddings, vector search, semantic similarity, and RAG?
Can you decide when prompt engineering, RAG, fine-tuning, agents, or guardrails are appropriate?
Can you choose the right evaluation metric for common scenarios?
Can you recognize responsible AI risks and mitigation controls?
Can you map IAM, KMS, CloudTrail, CloudWatch, Macie, Glue, S3, and Lake Formation to security and governance needs?

Practice plan after this Cheat Sheet

Use this Cheat Sheet as a checkpoint, then move into IT Mastery practice:

Start with topic drills on AI/ML fundamentals, generative AI, AWS service selection, responsible AI, and security.
Use original practice questions to force scenario recognition rather than memorization.
Read detailed explanations for every missed question and every guessed question.
Create a miss log with three columns: concept missed, why the wrong answer was tempting, and the decision rule to remember.
Take a mixed mock exam only after your topic drills show consistent performance across service selection, generative AI, evaluation, and governance.

Next step: choose a focused AIF-C01 question bank topic drill, answer without notes, then review the detailed explanations until you can explain why each wrong option is wrong.

Put the review into practice

Study Plan

AIF-C01 — AWS Certified AI Practitioner Cheat Sheet

Exam focus at a glance

Core AI, ML, and generative AI vocabulary

AWS AI service selection matrix

Managed AI services vs Amazon Bedrock vs Amazon SageMaker

AWS AI and ML services at a glance

Fast decision path: managed AI, generative AI, or custom ML

Amazon Bedrock and generative AI reference

Foundation model selection checklist

Prompt engineering quick reference

Prompt safety traps

RAG: Retrieval-Augmented Generation

RAG vs fine-tuning vs prompt-only

Machine learning problem types

Learning approaches

Amazon SageMaker lifecycle reference

Evaluation metrics

Confusion matrix terms

Classification metric selection

Regression metric selection

Generative AI evaluation

Foundation model concepts

RAG vs fine-tuning vs prompt engineering

Typical RAG flow

Classification metrics

Regression and generative AI evaluation

Responsible AI reference

Responsible AI principles to connect to scenarios

Responsible AI review

Security and governance for AI workloads on AWS

Shared responsibility reminder

Data foundations for AI on AWS

Deployment and inference patterns

Cost and performance levers

High-yield AWS distinctions

Common AIF-C01 scenario cues

Data quality and model quality traps

Quick pre-exam checklist

Exam mindset: how to choose the best answer

High-yield concepts to know cold

AI, ML, deep learning, and generative AI

Supervised, unsupervised, and reinforcement learning

Classification, regression, clustering, and anomaly detection

ML lifecycle quick review

Data concepts that commonly appear

Data types and storage patterns

Data quality and bias checks

Security, privacy, and governance decision rules

Core AWS controls

AI-specific security concerns

Cost, performance, and operational tradeoffs

High-yield scenario patterns

Common candidate mistakes

Fast final review checklist

Practice plan after this Cheat Sheet

Put the review into practice

Browse Certification Practice Tests