MockRoom Practice exams for cloud certifications
AIP-C01 · Professional

What the AWS Certified Generative AI Developer – Professional exam actually asks.

A professional-level certification for developers who build and run generative AI applications on AWS. This page covers the exam's shape, how its five domains are weighted, and how to practise for a long, scenario-heavy paper.

The exam at a glance

Figures from the AIP-C01 exam guide and the AWS certification page. AWS can change them, so check before you book.

  • 75 questions in 180 minutes — about 2 minutes 24 seconds each on an even split. The stems are long, so reading speed is part of the exam.
  • It sits at the Professional tier, the most advanced of AWS's role-based levels, and is aimed at people with real build experience.
  • 65 of the 75 are scored. The other 10 are unscored items AWS uses to trial future questions. You cannot tell which they are, and they do not count against you.
  • 750 out of 1000 to pass, on a scaled score. Because it is scaled, the number of questions you must get right is not fixed or published.
  • Answer everything. A blank scores nothing, so flag the long ones, bank the quick ones, and return before time runs out.

Three hours is a long time to read dense scenarios accurately. Pacing is a skill of its own: decide early how long a question may take before you flag it and move on.

Five domains, led by integration and data

Weightings are from the exam guide. They show where preparation time pays back most.

  • Domain 1 — Foundation Model Integration, Data Management, and Compliance (31%). The largest domain. Choosing and configuring foundation models for a requirement, preparing and validating the data they use, vector stores and retrieval for grounding answers, and managing prompts and the compliance obligations that come with the data.
  • Domain 2 — Implementation and Integration (26%). Building the application around the model: calling model APIs, agents and tool use, deploying models, and fitting generative AI into existing enterprise systems and workflows.
  • Domain 3 — AI Safety, Security, and Governance (20%). Controls on what goes into and comes out of a model, protecting data and access, and the governance and responsible-AI practices an organisation needs around a deployed system.
  • Domain 4 — Operational Efficiency and Optimization for GenAI Applications (12%). Controlling cost, latency, and throughput, and monitoring a generative AI application once it is live.
  • Domain 5 — Testing, Validation, and Troubleshooting (11%). Evaluating model and application quality, and diagnosing why a generative AI application is giving poor, slow, or inconsistent results.

Domains 1 and 2 together make up 57% of the exam. The two smallest domains are 23% between them, which is still roughly one question in four, so they are not safe to skip.

Scenario questions, and what they reward

Professional-level questions give a business situation with several competing constraints.

Expect a paragraph describing an application, its data, its users, and a constraint such as latency, cost, data residency, or auditability, followed by a question about the best change or design. More than one option may work; the exam asks for the one that satisfies every stated constraint. AWS publishes the response types it uses in the exam guide, so confirm them there before you sit the paper.

Read the last sentence first, then the scenario, then the options. Qualifiers to look for:

  • "Least operational overhead" → Prefer a managed capability over code you would have to build and maintain.
  • "Without sending data outside" a boundary → Check where the data goes, who can see it, and what is logged.
  • "Most cost-effective" → Look at how the model is invoked, how much context is sent each time, and whether results can be reused.

How to prepare for a professional exam

Reading about generative AI services is not enough at this level. The questions assume you have weighed trade-offs in a real design, so the study that pays is applying knowledge to scenarios.

Three habits that work:

  • Build something small end to end. A retrieval-grounded assistant with guardrails and some monitoring touches most of the domains and makes the questions concrete.
  • Sit full 180-minute timed mocks. Stamina and pacing across 75 long questions are things you can only train by doing them.
  • Review every option, not just your answer. Several options on this exam are workable; learning why one fails a specific constraint is the skill being tested.

Practise it free

No account, no paywall, no sign-up. Progress stays in your browser.

MockRoom's AIP-C01 study room has five distinct 75-question mock exams — 375 original scenario questions in total, written against the published exam guide and weighted across its five domains, each with an explanation for every option. It also has 100 active recall flashcards for spaced review. Nothing here is taken from an exam dump, and none of it is real examination content.

Open the AIP-C01 study room