AWS Certified AI Practitioner (AIF-C01)
By SpyderMan · Updated Aug 9, 2026
| Certification | AI Practitioner (AIF-C01) |
| Questions | 162 |
| Passing Score | 70% |
| Format | 80% Multiple Choice |
| Sessions Logged | 9,506 |
| Your progress | Log in / Register to track times taken, best score, questions mastered, and coverage on this quiz. |
| Rating |
Certification Summary
The AWS Certified AI Practitioner AIF-C01 is for cloud engineers, data analysts, and developers who manage model deployments within an existing infrastructure. You take this exam when your manager decides the team needs a generative interface for internal documentation and you need to understand the service constraints before the inevitable integration failure occurs.
The test balances model selection and training strategies, with Applications of Foundation Models, Fundamentals of GenAI, and Fundamentals of AI and ML accounting for nearly three-quarters of the content. The remaining questions address Guidelines for Responsible AI and Security, Compliance, and Governance for AI Solutions. You spend your time identifying which Amazon Bedrock configurations minimize data leakage while keeping latency within acceptable parameters for downstream service calls.
Topics covered
Fundamentals of AI and ML32 questions
- AI/ML concepts and terminology6 q
- ML problem types and use cases5 q
- ML lifecycle and data concepts5 q
- Inference, training, and evaluation concepts5 q
- AWS AI/ML services and use cases5 q
- Model selection and business considerations6 q
Fundamentals of Generative AI39 questions
- Generative AI concepts and terminology7 q
- Foundation models and model selection6 q
- Prompt engineering7 q
- Context, grounding, and RAG concepts5 q
- Generative AI applications and use cases5 q
- Model evaluation and output quality5 q
- AWS generative AI capabilities4 q
Applications of Foundation Models45 questions
- Foundation-model application architecture6 q
- Amazon Bedrock capabilities and use cases8 q
- Prompt engineering techniques6 q
- RAG and knowledge grounding6 q
- Fine-tuning/customization5 q
- Model evaluation and selection5 q
- Agents, orchestration, and application patterns5 q
- Generative AI application tradeoffs4 q
Guidelines for Responsible AI23 questions
- Responsible AI principles5 q
- Bias and fairness4 q
- Transparency and explainability4 q
- Privacy and data considerations3 q
- Safety, robustness, and human oversight4 q
- Responsible AI evaluation and governance3 q
Security, Compliance, and Governance for AI Solutions23 questions
- AI security concepts and threats5 q
- Data protection and privacy4 q
- AWS security controls for AI workloads4 q
- Compliance and regulatory considerations4 q
- Governance, risk, and auditability3 q
- Access control and operational safeguards3 q
Question catalog for this preview
Catalog listing of the 5 preview questions for this quiz.
Question 1
A fraud-detection API must return a risk score synchronously while a customer is waiting for a payment authorization decision. The request payload is small, and the response must be available within the transaction flow. Which inference pattern best fits the requirement?
Question 2
A support application has predictable, sustained model traffic and a strict throughput target throughout business hours. The team is comparing on-demand model invocation with Amazon Bedrock Provisioned Throughput. Which analysis is most important before choosing the provisioned option?
Question 3
Match each responsible-AI evaluation activity with what it is intended to reveal. Each response is used once; two responses are unused.
Question 4
A team is creating a prompt for an assistant that summarizes a document for a specific audience. The first version produces inconsistent results because it does not identify the task, audience, source boundaries, or output format. Which revision best addresses the missing prompt components?
Question 5
Match each AI safety or robustness concern with the most direct mitigation. Each response is used once; two responses are unused.