AIF-C01 · Exam scope · Services
In-Scope AWS Services
Services the AWS Certified AI Practitioner exam may reference—organized by category per AWS. List is non-exhaustive and subject to change.
Official list: In-Scope AWS Services (AIF-C01)
Exam guide hub: AWS Certified AI Practitioner (AIF-C01)
← → · Space · Home / End · Printable table: aif-c01-study-cheatsheet.html
How to use this deck
What the exam expects
Each category slide has a table: Service → Exam focus (what scenarios or distractors often test). For a dense all-in-one view, open aif-c01-study-cheatsheet.html in the same folder.
For these services, the AI Practitioner exam focuses on recognition, fit-for-purpose choice, and high-level integration—not deep hands-on implementation of every feature.
flowchart LR
S[Service name] --> U[Typical use case]
U --> G[Governance cost security fit]
Always verify the live AWS documentation list before you book—the exam guide is updated over time.
Definitions
- In scope
- Topics and services AWS may use when writing questions; being listed does not mean every feature is tested.
Analytics
Analytics
Service → typical exam scenario
| Service | Exam focus (what to pick it for) |
|---|
| AWS Data Exchange | Subscribe to third‑party datasets in AWS for analytics/ML without building custom data supply contracts/pipes in scenarios about licensed data. |
|---|
| Amazon EMR | Large‑scale distributed processing (Spark/Hadoop) for heavy data prep or feature engineering before training. |
|---|
| AWS Glue | Serverless ETL, crawlers, and Data Catalog—glue between S3/lakes and downstream ML/analytics. |
|---|
| AWS Glue DataBrew | Visual, low‑code data prep for analysts when questions emphasize no‑code cleaning/standardization. |
|---|
| AWS Lake Formation | Fine‑grained governance and permissions on a data lake for compliant ML/analytics access. |
|---|
| Amazon OpenSearch Service | Full‑text and k‑NN / vector search—classic distractor vs Kendra for “index/embed + similarity” RAG patterns. |
|---|
| Amazon QuickSight | BI dashboards and business metrics atop warehouses/lakes—when the stem asks for executive visibility, not model training. |
|---|
| Amazon Redshift | Cloud data warehouse/SQL analytics; training data or BI joins at warehouse scale. |
|---|
Exam trick: Match RAG retrieval to OpenSearch/vector DB options vs enterprise ML search (Kendra) per scenario wording.
Cloud Financial Management
Cloud Financial Management
Service → typical exam scenario
| Service | Exam focus |
|---|
| AWS Budgets | Set spend alerts/thresholds when scenarios mention controlling FM inference, training, or token burn vs plan. |
|---|
| AWS Cost Explorer | Analyze historical cost by service/tags/accounts when questions need visibility into what drove AI spend. |
|---|
Compute & Containers
Compute & containers
Service → typical exam scenario
| Service | Exam focus |
|---|
| Amazon EC2 | GPU/CPU instances for custom training, BYO containers, or legacy lift‑and‑shift model servers when “full control / specific AMIs” appears. |
|---|
| AWS Lambda | Short event-driven invocations: glue around Bedrock/SageMaker, file triggers, lightweight API proxies—contrasts with long‑running GPU training on EC2. |
|---|
| Amazon ECS | AWS‑native container orchestration for scalable inference microservices or batch workers without running Kubernetes. |
|---|
| Amazon EKS | Managed Kubernetes when portability, complex multi‑service ML platforms, or standard K8s tooling is required. |
|---|
Database
Database
Service → typical exam scenario
| Service | Exam focus |
|---|
| Amazon Aurora | Relational + often Postgres-compatible vector workloads; transactional data next to embeddings. |
|---|
| Amazon RDS | Managed relational; “vector on Postgres” pattern in RAG/knowledge scenarios (vs OpenSearch k‑NN). |
|---|
| Amazon DocumentDB | Mongo‑style JSON documents; scenarios with document + embedding storage when Mongo API is implied. |
|---|
| Amazon DynamoDB | Low‑latency key‑value for app metadata, sessions, feature lookups—not your first pick for heavy similarity search alone. |
|---|
| Amazon ElastiCache | In‑memory cache to cut latency or repeated inference/retrieval cost in high‑QPS designs. |
|---|
| Amazon Neptune | Graph store; graph + vector / relationship‑heavy knowledge when edges matter between entities. |
|---|
Developer Tools
Developer Tools
Service → typical exam scenario
| Service | Exam focus |
|---|
| Kiro | AWS builder/development experience (high‑level recognition)—pair with “accelerate coding with GenAI assistance” stems per current docs. |
|---|
| Strands Agents | Agent‑oriented builder tooling (recognition)—when answers reference agent development on AWS stack. |
|---|
| Amazon Q | GenAI assistant across IDE, console, or business apps—questions on “explain resource,” code help, internal doc Q&A. |
|---|
Names change—reconcile with the live in‑scope page before exam day.
Machine Learning · A–L
Machine Learning (part 1)
Service → typical exam scenario
| Service | Exam focus |
|---|
| Amazon A2I | Human review for low‑confidence predictions or regulated approvals—vs fully automated inference only. |
|---|
| Amazon Bedrock | Primary managed GenAI entry: FMs, Knowledge Bases, Guardrails, Agents—default answer for “invoke FM securely at scale.” |
|---|
| Amazon Bedrock AgentCore | Platform primitives for building production agents on Bedrock when stems mention agent runtime/orchestration. |
|---|
| Amazon Comprehend | Managed NLP: entity, sentiment, PII, classification—choose when input is raw text buckets, not generative authoring. |
|---|
| Amazon Kendra | Enterprise ML search with connectors—contrast with “build your own vector DB in OpenSearch” for curated enterprise Q&A. |
|---|
| Amazon Lex | Chat/voice bots with intents & slots; conversational IVR/support—contrast with open‑ended FM chat. |
|---|
| Amazon Nova | AWS foundation model family available via Bedrock ecosystem—recognition under FM selection questions. |
|---|
Machine Learning · M–Z
Machine Learning (part 2)
Service → typical exam scenario
| Service | Exam focus |
|---|
| Amazon Personalize | Real‑time recommendations (“customers who viewed”) vs training a bespoke FM from scratch. |
|---|
| Amazon Polly | Text‑to‑speech—voice output, IVR prompts, accessibility. |
|---|
| Amazon Rekognition | Image/video ML: labels, moderation, face detection workflows. |
|---|
| Amazon SageMaker AI | Full custom ML lifecycle (notebooks, training, endpoints, monitoring)—when you need BYO training, not only Bedrock APIs. |
|---|
| SageMaker JumpStart | Prebuilt models/solutions to start fast inside SageMaker vs coding pipelines from zero. |
|---|
| Amazon Textract | OCR + forms/tables from scans/PDFs—ingestion step before RAG or analytics. |
|---|
| Amazon Transcribe | Speech‑to‑text for calls, media captions, voice analytics pipelines. |
|---|
| Amazon Translate | Neural translation between languages in content/customer support scenarios. |
|---|
| AWS Transform | Modernization/transform programs to move legacy estates toward cloud‑ready analytics/AI consumption (recognition per guide). |
|---|
Management & Governance
Management and Governance
Service → typical exam scenario
| Service | Exam focus |
|---|
| AWS CloudTrail | Who invoked Bedrock / SageMaker / IAM changes—audit trail for security & forensics. |
|---|
| Amazon CloudWatch | Metrics, alarms, logs for latency/errors on endpoints, Lambdas, and operational health. |
|---|
| AWS Config | Record and rule‑check resource configurations (encryption on, tagging) for compliance drift. |
|---|
| AWS Trusted Advisor | Best‑practice checks (cost, security, limits)—“quick hygiene,” not full AI risk assessment. |
|---|
| AWS Well-Architected Tool | Structured reviews (incl. reliability/security pillars) for architecture documentation and gaps. |
|---|
Networking & CDN
Networking and Content Delivery
Service → typical exam scenario
| Service | Exam focus |
|---|
| Amazon VPC | Isolate subnets, security groups, private access to models/data; prerequisite language for “no public internet.” |
|---|
| Amazon CloudFront | CDN edge caching for global latency of apps, APIs, or static assets fronting AI experiences. |
|---|
Security, Identity, and Compliance
Security, Identity, and Compliance
Service → typical exam scenario
| Service | Exam focus |
|---|
| AWS IAM | Least‑privilege roles/policies for humans & services invoking models and reading training data. |
|---|
| AWS KMS | Customer‑managed keys to encrypt datasets, artifacts, and secrets—audit/regulated scenarios. |
|---|
| AWS Secrets Manager | Rotate & retrieve API keys/DB creds apps use next to GenAI pipelines. |
|---|
| Amazon Macie | Discover PII/sensitive data in S3—before fine‑tuning or indexing for RAG. |
|---|
| Amazon Inspector | Workloads vulnerability assessments where EC2/ECR workloads host ML services (recognition). |
|---|
| AWS Artifact | Download AWS compliance reports (SOC, etc.) for vendor due diligence. |
|---|
| AWS Audit Manager | Continuous evidence collection mapped to control frameworks for audits. |
|---|
Storage
Storage
Service → typical exam scenario
| Service | Exam focus |
|---|
| Amazon S3 | Central lake for raw data, curated training sets, model artifacts, inference logs—default object store in AI scenarios. |
|---|
| Amazon S3 Glacier | Long‑term archival/retention when “rarely accessed training history or compliance” is required. |
|---|
Also review
Out-of-scope & domain topics
The same exam guide includes an out-of-scope services list—review it so you do not over-study products AWS excludes from this exam.
Out-of-Scope AWS Services (AIF-C01)
Your Domain 1–5 slide decks explain concepts; this deck is a service checklist aligned to AWS’s published scope.
Cheatsheet
Printable study sheet
Open aif-c01-study-cheatsheet.html for all services + core terms in compact tables (browser Print → PDF optional).