AWS Generative AI Professional Exam Topics, Tips & Free Practice Exams

AWS Certified Generative AI Developer – Professional Exam Topics

Exam Basics

Exam code: AIP-C01
Format: 75 total questions, including 65 scored and 10 unscored questions
Question types: multiple choice and multiple response
Multiple-response questions require all correct responses for credit
Scoring: scaled from 100–1,000 with a passing score of 750
Scoring model: compensatory across the entire exam
Unanswered questions are incorrect, but there is no penalty for guessing
The exam validates production implementation of generative AI solutions on AWS

Target Candidate Experience

At least 2 years building production-grade applications on AWS or with open-source technologies
General AI/ML or data engineering experience
At least 1 year of hands-on experience implementing generative AI solutions
Experience with AWS compute, storage, databases, and networking
Understanding of AWS security, identity, deployment, infrastructure as code, monitoring, and cost optimization
Comfort with APIs, serverless services, containers, CI/CD, observability, and enterprise integrations

Domain 1: Foundation Model Integration, Data Management, and Compliance (31%)

Analyze business requirements and design GenAI architectures that fit technical, operational, and compliance constraints
Create proofs of concept with Amazon Bedrock and apply the AWS Well-Architected Generative AI Lens
Select foundation models by capability, limitations, benchmark results, cost, latency, and use-case fit
Design dynamic model selection, model fallback, graceful degradation, and Cross-Region Inference
Manage customized and fine-tuned models with SageMaker AI, Model Registry, versioning, rollback, and retirement processes
Validate and process text, image, audio, and tabular data for foundation model consumption
Use AWS Glue Data Quality, SageMaker Data Wrangler, SageMaker Processing, Amazon Transcribe, and Amazon Comprehend

Design vector stores with Bedrock Knowledge Bases, OpenSearch Service, Aurora PostgreSQL with pgvector, DynamoDB, RDS, and S3
Apply metadata, indexing, sharding, synchronization, refresh, and data-quality strategies
Select chunking, embedding, semantic search, keyword search, hybrid search, and reranking approaches
Use Amazon Titan embeddings and Bedrock reranker models
Implement query expansion, decomposition, transformation, function calling, and Model Context Protocol retrieval clients
Create governed prompts with Bedrock Prompt Management, Prompt Flows, reusable templates, approval workflows, and versioning
Test prompts for quality, edge cases, regression, structured outputs, context handling, and consistency

Domain 2: Implementation and Integration (26%)

Build agentic AI systems with memory, state, tools, stopping conditions, timeouts, and resource boundaries
Use Strands Agents, AWS Agent Squad, Model Context Protocol, Step Functions, Lambda, and IAM
Implement single-agent, multi-agent, ReAct, human-in-the-loop, and tool-calling patterns
Deploy models through Amazon Bedrock on-demand inference, provisioned throughput, SageMaker AI endpoints, and containers
Account for GPU, memory, model loading, token capacity, latency, throughput, and model size
Apply model cascading and use smaller specialized models for routine tasks

Integrate GenAI with enterprise applications through API Gateway, Lambda, EventBridge, SQS, Step Functions, and webhooks
Apply federation, role-based access, least privilege, hybrid integration, Outposts, and Wavelength
Build CI/CD pipelines with CodePipeline and CodeBuild, including tests, security scans, quality gates, and rollback
Implement synchronous, asynchronous, streaming, WebSocket, and server-sent event model APIs
Handle retries, exponential backoff, throttling, quotas, rate limits, fallbacks, and dynamic model routing
Build application interfaces with Amplify, OpenAPI, Bedrock Data Automation, Prompt Flows, and Amazon Q Developer

Domain 3: AI Safety, Security, and Governance (20%)

Filter unsafe or policy-violating inputs and outputs with Amazon Bedrock Guardrails
Build moderation workflows with Lambda, Step Functions, Amazon Comprehend, and custom classifiers
Reduce hallucinations through grounding, Bedrock Knowledge Bases, fact verification, confidence scoring, and structured outputs
Detect prompt injection, jailbreaks, adversarial inputs, unsafe tool calls, and model misuse
Implement defense in depth through preprocessing, model controls, post-processing, and API filtering
Protect deployments with VPC endpoints, IAM, least privilege, Lake Formation, KMS, and secure network boundaries

Detect and protect PII with Amazon Comprehend, Amazon Macie, masking, anonymization, redaction, and retention controls
Track data lineage, source attribution, metadata, prompts, outputs, and model decisions
Use AWS Glue Data Catalog, CloudTrail, CloudWatch Logs, model cards, and audit evidence
Monitor for misuse, drift, bias drift, policy violations, unsafe output, and compliance failures
Apply fairness evaluation, transparency, confidence, explanations, source attribution, and Bedrock agent tracing
Use LLM-as-a-judge, policy-based guardrails, model limitations, and automated compliance checks

Domain 4: Operational Efficiency and Optimization (12%)

Estimate, track, and reduce input and output token usage
Use context pruning, prompt compression, response limits, and conversation-history optimization
Apply model cascading, tiered model selection, batching, concurrency, and capacity planning
Optimize Bedrock provisioned throughput and resource utilization
Use semantic caching, prompt caching, result fingerprinting, deterministic hashing, and edge caching
Reduce latency through streaming, parallel requests, precomputation, and latency-optimized models
Optimize RAG indexes, queries, retrieval paths, and vector database performance
Tune temperature, top-k, top-p, batch inference, and concurrent model invocation
Monitor tokens, latency, throughput, cost, hallucination rate, drift, retrieval quality, and tool calls
Use Bedrock Model Invocation Logs, CloudWatch, X-Ray, Logs Insights, Synthetics, and Managed Grafana

Domain 5: Testing, Validation, and Troubleshooting (11%)

Evaluate relevance, factual accuracy, consistency, fluency, safety, usefulness, and business outcomes
Use Bedrock Model Evaluations, A/B tests, canary tests, multi-model comparisons, and LLM-as-a-judge
Collect human feedback, ratings, annotations, corrections, and user-experience signals
Build continuous evaluation, regression tests, automated quality gates, and deployment validation
Measure RAG relevance, context matching, retrieval latency, agent task completion, and tool effectiveness
Diagnose context-window overflow, truncation, missing context, and incomplete responses
Troubleshoot Bedrock API requests, credentials, permissions, quotas, parameters, and response handling
Correct prompt, chunking, embedding, vectorization, retrieval, drift, preprocessing, and output-schema issues
Use CloudWatch Logs, Logs Insights, X-Ray, synthetic workflows, output diffing, and golden datasets

Outside the Primary Exam Scope

Developing or training foundation models from scratch
Advanced machine learning research and algorithm development
Broad data engineering that is unrelated to GenAI application integration
Feature engineering for traditional machine learning workloads
Services and technologies that do not support the tested GenAI developer role

What the Exam Emphasizes

Production GenAI application design and implementation
Foundation model integration rather than model training
RAG, embeddings, vector stores, chunking, retrieval, and reranking
Prompt engineering, prompt governance, agents, tools, and MCP
Enterprise APIs, event-driven integration, serverless, containers, and CI/CD
Safety, privacy, security, governance, responsible AI, and auditability
Cost, token, latency, throughput, quality, and operational tradeoffs
Evaluation, observability, troubleshooting, and continuous improvement

Core Generative AI and Machine Learning Services

Amazon Bedrock, Bedrock AgentCore, Bedrock Knowledge Bases, Prompt Management, and Prompt Flows
Amazon SageMaker AI, Clarify, Data Wrangler, Ground Truth, JumpStart, Model Monitor, Model Registry, Neo, Processing, and Unified Studio
Amazon Titan models and embeddings
Amazon Comprehend, Kendra, Lex, Rekognition, Textract, and Transcribe
Amazon Q Business, Q Business Apps, Q Developer, and Amazon Quick
Amazon Augmented AI for human review workflows

Application, Data, and Integration Services

API Gateway, AppSync, Lambda, App Runner, EC2, Outposts, and Wavelength
Amazon ECS, EKS, ECR, and AWS Fargate
Amazon EventBridge, SNS, SQS, Step Functions, AppFlow, and AppConfig
Amazon OpenSearch Service, Aurora, RDS, DynamoDB, DynamoDB Streams, DocumentDB, Neptune, and ElastiCache
Amazon S3, EBS, EFS, Intelligent-Tiering, Lifecycle policies, and Cross-Region Replication
AWS Glue, Athena, EMR, Kinesis, MSK, and QuickSight

Security, Developer, and Operations Services

IAM, IAM Access Analyzer, IAM Identity Center, Cognito, KMS, Encryption SDK, Secrets Manager, Macie, and WAF
Amazon VPC, PrivateLink, CloudFront, Global Accelerator, Route 53, and Elastic Load Balancing
CloudTrail, CloudWatch, CloudWatch Logs, CloudWatch Synthetics, Managed Grafana, and X-Ray

AWS CDK, CloudFormation, CLI, CodeArtifact, CodeBuild, CodeDeploy, CodePipeline, Kiro, and AWS SDKs
AWS Auto Scaling, Systems Manager, Service Catalog, Well-Architected Tool, Cost Explorer, and Cost Anomaly Detection
AWS DataSync and Transfer Family for data movement and integration

How to Prepare for AIP-C01

Build RAG solutions with Bedrock Knowledge Bases, embeddings, vector stores, metadata filters, chunking, hybrid search, and reranking
Compare foundation models by quality, context window, latency, cost, modality, limitations, and business fit
Practice Prompt Management, Prompt Flows, Guardrails, prompt versioning, regression testing, and structured outputs
Build agents with tools, memory, state, MCP, Step Functions, Lambda, IAM restrictions, and human review
Implement Bedrock APIs with streaming, asynchronous processing, retries, throttling, routing, and graceful degradation
Practice enterprise integration with API Gateway, EventBridge, SQS, Step Functions, CodePipeline, and CodeBuild
Apply prompt-injection defenses, content moderation, PII detection, VPC endpoints, KMS, least privilege, and output filtering
Review model cards, lineage, audit logs, fairness, transparency, source attribution, responsible AI, and continuous governance
Optimize tokens, model selection, caching, provisioned throughput, latency, concurrency, RAG performance, and parameter settings
Evaluate and troubleshoot with Bedrock Model Evaluations, LLM-as-a-judge, human feedback, golden datasets, CloudWatch, X-Ray, and retrieval diagnostics

Top-Level AIP-C01 Exam Objectives

Foundation Model Integration, Data Management, and Compliance
Implementation and Integration
AI Safety, Security, and Governance
Operational Efficiency and Optimization for GenAI Applications
Testing, Validation, and Troubleshooting

AWS Generative AI Professional Practice Exams

AWS Generative AI Professional Practice Exams

AWS Generative AI Developer – Professional Exam Facts

AIP-C01 Content Domains & Weighting

The Trick to IT Certification Success

AWS Certified Generative AI Developer – Professional Exam Topics

Exam Basics

Target Candidate Experience

Domain 1: Foundation Model Integration, Data Management, and Compliance (31%)

Domain 2: Implementation and Integration (26%)

Domain 3: AI Safety, Security, and Governance (20%)

Domain 4: Operational Efficiency and Optimization (12%)

Domain 5: Testing, Validation, and Troubleshooting (11%)

Outside the Primary Exam Scope

What the Exam Emphasizes

Core Generative AI and Machine Learning Services

Application, Data, and Integration Services

Security, Developer, and Operations Services

How to Prepare for AIP-C01

Top-Level AIP-C01 Exam Objectives

Get AWS AI Practitioner Certified FAST!