Preparing for AI Certification: Topics, Practice, and Exam Procedures

Watch the step-by-step exam walkthrough

Follow 12 steps from sign-in and device checks to answering questions, submitting code, and finishing your exam.

Start with a competency map, not leaked questions

AI certification preparation is more focused when you know what is assessed. Open the program syllabus and note the topics you understand, those you find difficult, and the practice you need. Use three simple labels: can explain, can calculate or implement, and not yet able.

For Introduction to Artificial Intelligence, the published scope covers AI, machine learning and deep learning foundations; search and decision-making; probability and prediction evaluation; and algorithmic reasoning, ethics and limitations. Other programs have different scopes, so do not copy a study plan without adapting it.

Check your starting point without opening your notes. Explain a concept in your own words, work through a small calculation, then implement a simple function. Early mistakes help you set priorities; they do not reliably predict your exam score.

Practise four types of ability

1. Concepts and reasoning

Go beyond memorizing terminology. Explain how approaches differ, when a method makes sense, and its limitations. For example: why does model evaluation need data that was not used for training? Write a short answer, then compare it with your learning materials.

2. Manual calculations

Practise the arithmetic, probability and metrics in your syllabus. Use small numbers so you can check each step. An original practice example: a model makes 17 correct predictions out of 20. Its accuracy is 17 ÷ 20 × 100% = 85%. Next, explain why accuracy alone can be misleading with highly imbalanced classes. This is not an active exam question.

3. Coding and testing

For programs with Python, practise reading function contracts: inputs, outputs and boundary conditions. Test simple functions with typical inputs, empty inputs where allowed, and boundary values. Record the causes of errors and repeat the exercise without copying a solution. Explaining the logic matters more than memorizing one code pattern.

4. Case analysis

Practise choosing a course of action based on objectives, data and constraints. For instance, what should you investigate when a model performs well on training data but poorly on new data? List possible explanations and the evidence needed to distinguish them.

An example seven-day preparation plan

This plan is an example of how to organize your time, not a promise of readiness or passing in seven days. Allow more time if your initial self-assessment reveals gaps in your foundations.

Use the public interactive examples to become familiar with the exam experience. They do not represent the full difficulty range or question bank. Avoid sources offering leaked questions, and never share active exam questions.

A checklist before selecting Start

• Access: confirm the program, registered email, schedule and exam code.
• Devices: connect your laptop or desktop to power, check the camera, and enable microphone permissions.
• Audio: make a short test recording and play it back in the device check.
• Environment: use adequate lighting, keep your face visible, check your connection, and minimize interruptions.
• Supplies: prepare blank paper and a pen for manual calculations.
• Rules: read the supervision policy, participant terms and batch instructions.

Checking devices before the exam does not use an exam attempt. One authorization grants one attempt; refreshing resumes the same session without resetting its timer. Contact the organizer before starting if you need an accommodation or encounter a problem.

What to do on exam day

Check the access opening time and deadline in your dashboard. For paid public registration, the exam information states a deadline of 23:59 WIB on your selected date. Starting close to that deadline can shorten the time available, even when the program’s standard duration is 120 minutes.

Read each section’s instructions and allocate time according to the number of tasks and their difficulty for you. In coding sections, Run and Submit do different things. Run checks public examples; Submit sends your solution for grading. If you edit your code after submitting, submit it again so the latest version is graded.

Work independently. Calculators, AI chatbots, LLM assistance and help from others are prohibited under the general policy unless an explicitly approved, applicable accommodation allows otherwise. Tabs or windows may be used to test your own code under the rules, but switches and time away are recorded. Permission to switch tabs is not permission to use prohibited assistance.

If your connection drops, try to resume the same session and report the issue through support. Do not assume the timer has stopped or extra time will automatically be granted. Record the time and a description of the problem without taking screenshots of exam questions.

After submission: review results and follow up

Check your section scores to identify areas for further study. The general passing standard is 75/100, subject to the applicable session or batch requirements. Certificates also require integrity-review approval.

If you dispute a result or supervision decision, use the appeal route in your dashboard and explain the basis of your appeal. Do not start a new session assuming your previous attempt has automatically been restored. A retake requires a new authorization under the applicable terms.

Once your certificate is issued, check its verification record before sharing it. Combine assessment results with projects and an explanation of your learning process to give a fuller picture of your abilities.

Frequently asked questions