Passing the AWS Certified AI Practitioner Foundational Certification (AIF-C01) in 2025
The AWS Certified AI Practitioner Foundational Certification will help validate in-demand knowledge of artificial intelligence (AI), machine learning (ML) and generative AI concepts and use cases. This certification helps you learn about how to approach this topic in the AWS ecosystem.
As stated in the AWS Certified AI Practitioner (AIF-C01) Exam Guide, this certification covers the following domains with the accompanying weightings:
- Domain 1: Fundamentals of AI and ML (20% of scored content)
- Domain 2: Fundamentals of Generative AI (24% of scored content)
- Domain 3: Applications of Foundation Models (28% of scored content)
- Domain 4: Guidelines for Responsible AI (14% of scored content)
- Domain 5: Security, Compliance, and Governance for AI Solutions (14% of scored content)
Study Approach
The AWS Certified AI Practitioner site provides a good overview of the certification material. You can find detailed information inside the Exam Guide as well. For most certifications, I start with a video course to get started. For this certification, I would suggest you try one of the following:
- Stephane Maarek’s Ultimate AWS Certified AI Practitioner AI-C01 course (Get a coupon for Stephane Maarek’s course at Datacumulus ).
- Frank Kane’s Mastering AWS Certified AI Practitioner AIF-C01 — Hands On! (Get a coupon from his Sundog Education website https://www.sundog-education.com/ )
You should be able to get these courses for somewhere around $20. Once you have gone through a video course, you should have a good understanding of the overall material and have a basic understanding of what is required.
Practice Questions
Now, getting ready for any certification involves some form of practice question training. I try to get the best set of questions based on recent reviews and prefer to have multiple question sets to train on.
I used the following AI Practitioner Practice tests
- Tutorials Dojo AWS Certified AI Practitioner Practice Exams AIF-C01 (135 Questions, 2 Full Tests and 5 extra, with randomized full tests)
- Whizlabs Certified AI Practitioner Practice Tests (130 questions, 2 full tests)
- ExamTopics Amazon AWS Certified AI Practitioner AIF-C01 Exam (20 free questions, 85 overall)
- AWS Certified AI Practitioner Official Practice Question Set (AIF-C01) (20 free questions)
Other question sets you might consider:
- AWS Certified AI Practitioner — AIF-C01 (Stephane Maarek & Abishek Singh) (260 questions (4x65)) (Check https://datacumulus.com/ for coupon)
- Practice Exams | AWS Certfied AI Practitioner (Thomas Hass) (65 Questions)
Practice Test Training
Now preparing for the examination involves doing plenty of training on the practice tests above. This will help you identify your areas of weakness and areas you need to dive into a bit more. Using multiple question materials will benefit you as they are all a bit different and emphasize different areas.
There are 3 main sets of questions I used and I will discuss them below. Generally, I like to get to a point where I would get a 100% on exams just to be sure I had a good handle on the testing material. The testing material is not always the same as the actual exam but maximizing your mastery of the testing material should help with the real exam.
Tutorials Dojo offers good questions to practice on and you can review your incorrect responses to get a better grasp of the material. There are different modes as well such as Timed Mode and Review Mode to change things up a bit. The final mode is the “Final Test” which offers a randomized set of 65 questions.
You will start out completing the tests taking a fair amount of time and with less than desired marks. If you keep practicing you should see improvement to high marks and the testing time will come down drastically.
When I can get a 100% on a fully randomized test and average near that I generally feel ready to take the real test. Coincidentally, this is also the point where I am also getting really tired of studying as well.
** Note the 47.37% was some kind of issue they had with their website in marking. Definitely an annoying thing but it happened several times in the last week when I was doing fully randomized question sets
Good set of questions that are really indicative of the actual exam questions. There is a bunch of annoying CAPTCHA and password items when using the free version which is understandable.
You have to discern between the “Most Voted” answer and the initial answer that ExamTopics presents. Generally, I would go with the answer that was selected based on the Community vote distribution after reading the comments. There is a Discussion popup and it will show you other viewers comments and generally the “Most Voted” response is the one you should really consider in my opinion. You can also open up the Discussion to see the arguments for the different choices.
Exam Topics does not actually provide 65 question sets and mark them. There currently are not a lot of questions there but they are very realistic
Whizlabs offers additional questions at a reasonable price. Questions are not randomized for most of the questions. This tends to lead you to actually start to remember the right responses based on memory.
Personal Tips on questions
Be consistent in taking your tests to get more comfortable with the material. Try to figure out what is the best time of day for you to practice and try to set up a schedule to practice.
Also, don’t be afraid to get poor marks at the start. View that as an opportunity to improve and each time you take a new test after reviewing the incorrect responses from last time. You will quickly notice that your scores will get better and the time to complete a test will come down as well.
Key Points
A2I (Amazon Augmented AI) — makes it easier to incorporate developer reviews of ML predictions, removing the need to build human review systems or manage large numbers of human analysts.
A/B testing is a controlled experiment, and a widely adopted practice in the tech industry. It involves simultaneously deploying multiple variants of a product or feature to distinct user segments
AI Service Cards — a form of responsible AI documentation that provide customers with a single place to find information on the intended use cases and limitations, responsible AI design choices, and deployment and performance optimization best practices for our AI services
Association rule learning — uncover rule-based relationships between inputs in a dataset
Asynchronous inference — ideal for requests with large payload sizes (up to 1GB), long processing times (up to one hour), and near real-time latency requirements
AUC — The area under the curve (AUC) metric is used to compare and evaluate binary classification by algorithms that return probabilities, such as logistic regression.
AWS Bedrock Model Evaluation — allow you to compare model or inference profile outputs, and then choose the model best suited for your downstream generative AI applications.
Amazon Lex is a service that allows developers to create conversational interfaces using voice and text. Also allows developers to create conversational interfaces using voice and text and involves a multi-channel approach.
Bedrock Model Evaluation — enables you to assess, compare, and choose the most suitable foundational models for your specific needs
Clustering — finding patterns or groups in large datasets that have not been explicitly labeled
Core Dimensions of Responsible AI
- Fairness — Considering impacts on different groups of stakeholders
- Explainability — Understanding and evaluating system outputs
- Privacy and Security — Appropriately obtaining, using, and protecting data and models
- Safety — Preventing harmful system output and misuse
- Controllability — Having mechanisms to monitor and steer AI system behavior
- Veracity and robustness — Achieving correct system outputs, even with unexpected or adversarial inputs
- Governance — Incorporating best practices into the AI supply chain, including providers and deployers
- Transparency — Enabling stakeholders to make informed choices about their engagement with an AI system
Context window — the amount of text an AI model can handle and respond to at once; in most LLMs, this text is measured in tokens.
Data Wrangler — reduces data prep time for tabular, image, and text data from weeks to minutes
Domain adaptation fine-tuning allows you to leverage pre-trained foundation models and adapt them to specific tasks using limited domain-specific data.
Epoch Count — determines how many times the model goes through the entire training dataset.
Few-shot prompting — to provide a few examples to help LLMs better calibrate their output to meet your expectations
Zero-shot prompting — where no example input-output pair is provided in the prompt text
Generative Adversarial Networks (GANs) are a type of machine learning model designed to generate new data by learning from an existing dataset.
Hallucination — responses that are not grounded in enterprise data or are irrelevant to the users’ query.
Hyper parameter tuning techniques include Bayeseian optimization, Grid search and Random Search
Instruction-based fine-tuning — is a process where a pre-trained foundation model is further trained with specific instructions to perform particular tasks
Amazon Kendra — finds answers faster with intelligent enterprise search powered by machine learning
Knowledge cutoff — when models providing outdated information, which is problematic in dynamic environments where up-to-date knowledge is crucial
Model Parallelism — a distributed training method in which the deep learning model is partitioned across multiple devices, within or across instances.
Model Theft — where a threat actor attempts to replicate a machine learning model without internal access to the systems of data.
Partial Dependence Plots (PDPs) are a powerful tool for explaining machine learning models by showing the relationship between features and the model’s predictions.
Probability density -To estimate how users’ preferences are spread across different genres
Prompt leaking — is a special type of injection that not only prompts the model to override instructions, but also reveal its prompt template and instructions
Prompt injection — are attacks involving manipulating prompts to influence LLM outputs, with the intent to introduce biases or harmful outcomes.
N-gram — contiguous sequences of n items from a given sample of text or speech
Red teaming — engages human testers to probe an AI system for flaws in an adversarial style, and complements our other testing techniques
Reinforcement learning — a machine learning (ML) technique that trains software to make decisions to achieve the most optimal results.
Rekognition Content Moderation Feature — automates and streamlines your image and video moderation workflows using machine learning (ML), without requiring ML experience.
Root mean squared error (RMSE) — used for regression models. Root mean squared error, or the standard deviation of the errors.
Sagemaker Clarify — helps identify potential bias during data preparation without writing code
Amazon SageMaker Feature Store — is a fully managed service that allows organizations to centrally store, share, and manage machine learning features across multiple projects
AWS Sagemaker Jumpstart — Machine learning (ML) hub with foundation models, built-in algorithms, and prebuilt ML solutions that you can deploy with just a few clicks
Sagemaker Model Cards — assist in managing and documenting the lifecycle of your machine-learning models.
Softmax layer — data scientists often use the softmax probability score as a notion of confidence about whether or not a predicted class is correct:
Supervised NLP — methods train the software with a set of labeled or known input and output.
Training Data Poisoning — occurs when ML models are trained on tampered data, leading to inaccurate model predictions
Trainium is a custom-built silicon by AWS optimized for deep learning workloads, making it cost-efficient and scalable for training large and complex models.
Transfer learning — involves taking a pre-trained model, which has been trained on a large dataset, and adapting it to a new, related task
Word2vec algorithm helps capture semantic and syntactic relationships between words
Evaluation Metrics
ROUGE (Recall-Oriented Understudy for Gisting Evaluation) is a set of metrics specifically designed to evaluate the quality of text summaries by comparing them to human reference summaries
BERTScore is a tool that compares how similar generated text is to a reference by understanding the context of words leveraging BERT (Bidirectional Encoder Representations from Transformers) embeddings
Bidirectional Encoder Representations from Transformers (BERT) — a bidirectional model, examines the context of an entire sequence before making predictions including missing words
Bilingual Evaluation Understudy (BLEU) is a metric that is specifically designed for evaluating machine translation.
Accuracy is a commonly used metric to evaluate the performance of classification models in machine learning
Recall -measures the proportion of actual positive instances (true positives) correctly identified by the model.
Precision is incorrect because it is a metric that measures the proportion of correct predicted positive instances
Exam Tips
Transfer Learning vs. Domain Adaption fine-tuning — A subtle key distinction here is that Domain Adaption uses limited domain specific data
Transparency vs. Explainability — transparency offers detailed breakdown, explainability provides steps on how the decision was arrived at or details on each decision made
Accuracy is the answer for determine how well the model correctly classifies
Using labels is always Supervised learning, no labels is unsupervised learning
Ground Truth involved human labelling
ROUGE is for summarization or summaries
Questions and Topics not seen in the training materials (You should review these to go above and beyond the training materials)
I have presented these that you can consider above and beyond the training materials mentioned previously
Know how whether to apply Supervised or Unsupervised learning techniques to each of the following:
- Binary classification (Supervised)
- Multiclass classification (Supervised)
- K-nearest neighbors algorithm (Supervised)
- Density estimation (Unsupervised)
Know how temperature affects the consistency of responses.
A lower temperature influences the model to select higher-probability outputs and higher temperature influences the model to select lower-probability outputs.
So I believe that a low temperature will improve the consistency
Know the features OpenSearch when used as a vector database
- Semantic search
- Retrieval Augmented Generation with LLMs
- Recommendation engine
- Media search
Amazon Bedrock Knowledge Bases — can give FMs and agents contextual information from your company’s private data sources for RAG to deliver more relevant, accurate, and customized responses
Conclusion
You will learn a lot about AWS Artificial Intelligence and machine learning by studying for this certification. Use this as an opportunity to learn more. Use a video course and practice often to prepare sufficiently before taking the test.
Best of luck on your preparation and hopefully this article helps you to pass the AWS Certified AI Practitioner Foundational Certification
Best of luck!
