AIF-C01 Study Plan: 4-Week and 8-Week

The short version

Most people with some AWS or IT background can prepare in four weeks at about 5 to 8 hours a week. If AI and the cloud are both new to you, eight weeks at 3 to 5 hours a week is calmer. These hours are rough guesses from general experience. They are not figures from AWS, and your pace will differ.

How the plan is built

The plan follows the domain weightings in AWS exam guide version 1.2 (unchanged from version 1.1): Applications of Foundation Models 28%, Fundamentals of GenAI 24%, Fundamentals of AI and ML 20%, and Responsible AI and Security/Governance 14% each. That means about half your time goes to the two generative AI domains.

AWS offers a free exam prep plan for AIF-C01 on Skill Builder, which you can use as your main learning material: AWS Skill Builder exam prep plan. Whatever you choose as your main source, keep the AWS exam guide open as your checklist and tick off each objective. Our general advice is on how to study for IT certification exams.

The 4-week plan

This plan suits you if you already know basic cloud ideas such as IAM and S3.

Week 1: AI and ML fundamentals (Domain 1). Learn the vocabulary: AI, ML, deep learning, training, inference, bias, and fit. Learn supervised, unsupervised, and reinforcement learning, and the inference types (batch, real-time, asynchronous, serverless). Match each managed service to its job: Transcribe, Translate, Comprehend, Lex, Polly, Rekognition, Textract. Learn precision, recall, and F1. Take a short quiz on the vocabulary at the end of the week.

Week 2: Generative AI and foundation models, part 1 (Domain 2 and the start of Domain 3). Cover tokens, embeddings, chunking, transformers, diffusion, and the foundation model lifecycle. Learn what hallucination is and why output can vary. Get to know Amazon Bedrock, SageMaker AI and JumpStart, and the idea of agents and tools.

Week 3: Applications of foundation models (Domain 3). This is the heaviest domain. Work through inference parameters, RAG and vector stores, the cost ladder of prompting, RAG, fine-tuning, and pre-training, prompt techniques (zero-shot, few-shot, chain-of-thought), and the evaluation metrics ROUGE, BLEU, and BERTScore. Take your first full practice set at the end of the week and read every explanation.

Week 4: Responsible AI, security, and review (Domains 4 and 5). Cover fairness, bias, overfitting and underfitting, transparency versus explainability, Bedrock Guardrails, model cards, IAM roles, encryption, Macie, PrivateLink, prompt injection, and governance services. Then use your practice results to revisit your weakest domain. Take a second practice set mid-week and a timed one near the end.

The 8-week plan

This plan suits you if cloud, or AI, or both are new.

  1. Week 1: Cloud basics: the shared responsibility model, IAM, S3, EC2, Lambda, and how AWS pricing works. Skip this if you are already comfortable.
  2. Week 2: Domain 1 part one: AI, ML, and deep learning, data types, learning types.
  3. Week 3: Domain 1 part two: use cases, when AI is not the right tool, the ML lifecycle, metrics, and the managed AI services.
  4. Week 4: Domain 2: generative AI concepts, strengths and limits, Bedrock, SageMaker, cost basics, and agent concepts.
  5. Week 5: Domain 3 part one: model selection, inference parameters, RAG, and embeddings. First practice set.
  6. Week 6: Domain 3 part two: prompt engineering, customization, fine-tuning data, and evaluation. Second practice set.
  7. Week 7: Domains 4 and 5: responsible AI, security, and governance. Third practice set across all domains.
  8. Week 8: Review weak areas, redo flagged questions, one timed set, and the final 3 days below.

Using practice exams in the plan

  • Take one set cold early on, only to see your baseline by domain. A low score here is useful, not bad.
  • Read the explanation for every question, including those you got right. On this exam the wrong options are usually real services, so the explanation teaches you the boundary between neighbors.
  • Keep a short list of confusions, such as “Comprehend or Textract?” and review it every few days.
  • Take at least one set under the real time limit of 90 minutes for 65 questions, with no pausing and no looking things up.
  • Do not memorize question wording. Questions on the real exam will be phrased differently.

The last 3 days

  • Day 3 before: Take one timed set. Review only the domains where you lost the most points.
  • Day 2 before: Reread your confusion list and skim the exam guide objectives. Stop learning new topics.
  • Day 1 before: Light review for an hour at most. Check your booking, ID, and, for an online exam, your room and computer. See the exam day guide.
  • Exam day: Eat, arrive early, flag long questions and return to them. Answer everything.

More guides for AIF-C01

Next step

Read domains explained to turn each week above into topics. When you want practice material, see our AIF-C01 practice tests.

Frequently asked questions

How many hours should I plan?

A common range is 20 to 40 hours in total for people with some background, and more if the field is new. This is general experience, not an AWS figure.

Should I do hands-on labs?

They are not required, since the exam is conceptual. A few minutes in the Amazon Bedrock console helps ideas stick, but watch for any usage costs on your own account.

What if my practice scores stay low?

Look at the domain breakdown rather than the total. Revisit the weakest domain and retake a set a week later.

Try these practice exams