← Learn
Learn · Cloud

AWS generative AI developer exam prep

Seventeen lessons for the AWS Certified Generative AI Developer – Professional exam (AIP-C01), following the five domains of AWS’s exam guide: calling and choosing models in Amazon Bedrock, preparing data, vector stores and retrieval, prompts, agents and MCP, deployment and integration, guardrails, security and governance, cost and monitoring, and evaluation and troubleshooting. The practice questions are scenarios in the exam’s style: a company, a need and a constraint, and the one answer that fits best.

An unofficial study course, not made by or affiliated with Amazon Web Services.

  1. The exam and how to read a question

    What the AIP-C01 exam covers, how it's scored, and how to find the one answer a long scenario is really asking for.

    7 questions

  2. Calling a model in Amazon Bedrock

    InvokeModel and the Converse API, messages and system prompts, streaming, and the parameters that shape a reply.

    7 questions

  3. Choosing, switching and customizing models

    Picking a model for the job, routing between models, cross-Region inference, fallbacks, and when to fine-tune.

    7 questions

  4. Preparing data for a model

    Validating, cleaning and converting text, documents, audio and images into input a model can use well.

    7 questions

  5. Embeddings and vector stores

    Turning text into vectors, choosing where to keep them, adding metadata, and keeping the index up to date.

    7 questions

  6. Chunking, search and reranking

    Splitting documents well, hybrid search, rerankers, query rewriting, and the Retrieve and RetrieveAndGenerate APIs.

    7 questions

  7. Prompt engineering and prompt management

    Writing prompts that hold up, managing them as versioned templates, chaining them in flows, and testing them.

    7 questions

  8. Agents, tools and MCP

    Bedrock Agents, AgentCore, Strands Agents, tool definitions, MCP servers, multi-agent setups and keeping agents in bounds.

    7 questions

  9. Deploying and serving models

    On-demand, batch and provisioned throughput in Bedrock, SageMaker AI endpoint types, and serving large models.

    7 questions

  10. APIs and enterprise integration

    API Gateway, streaming, queues and events, retries and throttling, CI/CD, and fitting AI into existing systems.

    7 questions

  11. Guardrails and content safety

    Amazon Bedrock Guardrails policies, prompt attacks, grounding checks, ApplyGuardrail, and defense in depth.

    7 questions

  12. Data security and privacy

    Private networking, least-privilege IAM, encryption, what Bedrock does with your data, PII detection and retention.

    7 questions

  13. Governance and responsible AI

    Audit trails, invocation logs, lineage, model cards, fairness and bias checks, transparency, and policy compliance.

    7 questions

  14. Cost and performance

    Spending fewer tokens, caching prompts and answers, picking cheaper models, and cutting latency.

    7 questions

  15. Monitoring and observability

    Bedrock metrics in CloudWatch, invocation logs, tracing, dashboards, alarms and cost anomaly detection.

    7 questions

  16. Evaluating models, RAG and agents

    Bedrock evaluations, LLM-as-a-judge, RAG metrics, human review, A/B and canary tests, and quality gates.

    7 questions

  17. Troubleshooting GenAI apps

    Reading the errors, context overflow, cut-off answers, bad retrieval, prompt drift and format failures.

    7 questions
More info

How it works

Each lesson is a short read on one part of the exam guide, then practice questions written like the exam’s: a short scenario with four options, where more than one could work and the constraint (cost, latency, least effort, compliance) decides which is best. The explanation says why. A few ask you to type a name or a number, or tap steps into order. You can try again as often as you like; after two misses there’s a button to show the answer.

What do I need to know first?

It's a professional-level exam, so it helps to know AWS basics (IAM, Lambda, S3, API Gateway, CloudWatch) and what a foundation model is. Each lesson explains the AI ideas it uses, so it works on its own.

Related courses

Where is my progress saved? In this browser, like the rest of the site. Nothing to sign up for.

Is it up to date? It follows AWS’s exam guide for AIP-C01 as published in October 2026. AWS services change often, so the lessons leave out prices and most quotas; check the current AWS documentation and exam guide before your exam.

An unofficial study course. Not made by, endorsed by or affiliated with Amazon Web Services; the questions are our own, not from the exam, and finishing it is not an AWS certification. AWS, Amazon Bedrock and the other service names are trademarks of Amazon.com, Inc. or its affiliates.