I believe many have heard about this certification by AWS. I have read that the AWS Certified Machine Learning – Specialty is a tough certification exam, and I couldn’t agree more.
I sat for this exam on Oct 1, 2021. Upon reading and answering the first few exam questions, I immediately sensed that it would not be easy to pass. But don’t be dismayed, because you definitely can do it if you put in the effort and make ample preparations. Believe in yourself, hard work does pay off.
Going through and passing this exam has been a great learning journey for me.
I’m glad that I managed to reach this milestone. It’s both satisfying and rewarding.
About the Exam
"The AWS Certified Machine Learning – Specialty Certification is intended for individuals who perform a development or data science role. It validates a candidate’s ability to design, implement, deploy, and maintain machine learning (ML) solutions for given business problems" – Source: AWS
The exam duration is 3 hours. There are a total of 65 questions, and that means you have approximately 2.7 minutes to answer each question on average. You will get either multiple choice or multiple response questions.
Out of the 65 questions, there are 15 unscored questions. AWS uses them to evaluate candidate performance and whether to include them as scored questions in the future. Unfortunately, they are not disclosed on the exam.
Results are reported as a scaled score of 100 to 1,000. The minimum passing score is 750.
The cost for this exam is 300 USD. If you have previously passed an AWS certification exam, you can utilize a 50% discount voucher from your AWS certification account.
Test Domains
There are 4 domains that the candidate would be tested on. The domains and their weightings are summarized in the figure below.

For the full detail and scope of each domain, you can read them in the exam guide that you can download from AWS. Below is a brief summary:
- Data Engineering – know the storage mediums (S3, EFS, FSx, etc), data ingestion (Kinesis, MSK, Firehose, etc), data transformation (ETL, Glue, EMR, AWS Batch, Lambda)
- Exploratory Data Analysis -covers data preparation (Ground Truth, Data Wrangler, etc), feature extraction and engineering (OHE, PCA, etc), analysis and visualization (correlation, residual plot, etc), Feature Store,
- Modeling – know the various algorithms related to supervised/unsupervised learning, transfer learning, reinforcement learning, Machine Learning types (classification, regression, forecasting, clustering, recommendation), optimization, hyperparameters tuning, regularizations (L1/L2, dropout), model evaluation (underfitting/overfitting), metrics (accuracy, precision, recall, RMSE, F1 score, AUC-ROC), confusion matrix, ML experiment, debug, detect bias & explain model predictions (SageMaker Experiment, Debugger, Clarify), etc.
- Machine Learning Implementation & Operations – monitoring and logging (Model Monitor, CloudTrail, CloudWatch), ML & AI Services (Lex, Comprehend, Transcribe, Rekognition, Textract, etc), IAM, Security, Endpoints with API Gateway, batch transform, A/B testing, BYOC, IoT Greengrass, SageMaker Neo, Augmented AI, etc.
Learning Options
There are many learning options that you can take prior to the exam. I have laid down the available possibilities in the figure below.

Depending on your level of experience and knowledge, you may or may not need all of them. I will explain each of these in the next section.
📑 1. AWS Sample Questions
First of all, I would recommend you download the sample questions from AWS. It consists of 10 questions with answers. The explanation and rationale behind each answer are also provided.
Try to answer these questions without looking at the answer. Then, check your answers against the answers provided. It should give you a feel of how hard the level of questions is. It also enables you to roughly gauge where you stand, whether you’re doing fine or need further training and studies.
💡 2. Machine Learning Certification Courses
If you feel that you need more training and studies, there are several AWS machine learning certification courses that you can find on Udemy, Whizlabs, A Cloud Guru, etc.
I took the Udemy course AWS Certified Machine Learning Specialty (MLS-C01) by Chandra Lingam. It’s a hands-on AWS SageMaker course with Practice Test included. On top of that, it also covers a lot of fundamental machine learning concepts and important algorithms.
I am impressed with the quality of this course. I particularly love the many hands-on labs and use cases presented to demonstrate specific concepts or implementations. The instructor also actively maintains and keeps the course materials up-to-date. I completed the course in late March but delayed my plan to take the exam. When I went through the course again in September, I was surprised that new topics had been added, which I found all relevant and useful.
According to students, the instructor is highly responsive and helpful. There are even organized group live Q&A sessions. In short, this is an awesome course not to be missed.
📚 3. AWS Resources and Documentation
There are a lot of supporting learning materials and resources that you can find from AWS. AWS maintains extensive documentation for all its services and products. I provide all the main links below.
👉 **** Machine Learning on AWS
This is the main page of AWS machine learning and it contains all the links to AWS AI Services, ML Services, Frameworks, Infrastructure, Use Case Solutions, Blog, learning resources, customer stories, etc.
The Amazon Machine Learning Developer Guide is a good resource explaining machine learning concepts and guides.
👉 **** AWS Documentation
Here, you can find user guides, developer guides, API references, tutorials and projects, SDKs, and toolkits.
The Amazon SageMaker Developer Guide is particularly useful. It contains very detailed information and tutorials on SageMaker features, as well as all the built-in algorithms.
👉 AWS Whitepapers
- Machine Learning Best Practices for Public Sector Organizations
- Machine Learning Lens – AWS Well-Architected Framework
- Augmented AI: The Power of Human and Machine
- Storage Best Practices for Data and Analytics Applications
👉 **** Amazon SageMaker Technical Deep Dive Series
There are now 20 videos in this Amazon SageMaker Technical Deep Dive Series on YouTube. Learn how to build, train, tune, deploy and optimize models with Amazon SageMaker.
👉 **** AWS Skill Builder
Learn in your own time and schedule with digital training on AWS Skill Builder. Browse and choose relevant courses from the learning library.
👉 **** AWS Product & Technical FAQs
FAQs are another great resource for skimming through the main points and features of a specific product or service.
Just to list a few: Amazon SageMaker FAQ, Amazon S3 FAQ, AWS Glue FAQ, Amazon Kinesis Data Streams FAQ, Amazon Data Firehose FAQ, AWS Lake Formation FAQ, Amazon Comprehend FAQ, Amazon Rekognition FAQ, etc.
👉 AWS Prescriptive Guidance Patterns
This guidance provides step-by-step instructions, architecture, tools, and code for implementing specific cloud migration, modernization, and deployment scenarios. You can find the machine learning & AI topics here.
💻 4. Practical Labs & Exercises
If you do not have experience with AWS, try to work on hands-on labs in AWS such as data preparation and transformation, and building, tuning, and deploying models.
This move is optional but it has the added advantage of helping you gain familiarity with AWS services and products, and how they integrate with each other.
Sign up for an AWS account and you will be able to explore, experience, and start building on AWS. New account holders can enjoy most of the products and services under [Free Tier](https://aws.amazon.com/free/). Check out AWS’s Free Tier offerings, which can fall under one of three categories: short-term free trials, 12 months free, or always free.
Links to example notebooks, SDKs, AWS CLI and AWS CDK are provided below:
- Amazon SageMaker Example Notebooks
- Hands-on Tutorials
- Amazon SageMaker Python SDK is an open-source library for training and deploying machine-learned models on Amazon SageMaker.
- [New] SageMaker-Core was introduced in early Sep 2024 as a Python SDK that provides an object-oriented interface for interacting with SageMaker resources such as TrainingJob, Model, and Endpoint resource classes. Key features include resource chaining, intelligent defaults, and logging configuration. For example notebooks, check out the sagemaker-core repository.
- Boto3 is AWS SDK for Python that you can use to create, configure, and manage AWS services.
- AWS Command Line Interface (AWS CLI) is a unified tool that provides a consistent interface for interacting with all parts of Amazon Web Services.
- AWS Cloud Development Kit (AWS CDK) is a framework for defining and provisioning cloud infrastructure in code.
- AWS Solutions Constructs (Constructs) is an extension of the AWS CDK that provides pre-built, well-architected patterns for quickly defining solutions in code to create predictable and repeatable infrastructure.
If you’re interested, you can click the link below to read the article I’ve written on how to build and train a recommender engine with Amazon SageMaker Factorization Machines.
📝 Take Your Own Notes
As the scope of coverage for this exam is quite enormous, I recommend taking your own notes while you go through learning materials.

You may not feel the need at the beginning. As you progress and dig through chapters of studies and check into piles of documentation and readings, you will soon find yourself overwhelmed with information overload, failing to remember features or stuff that you’ve studied earlier.

If you’re taking courses, they usually come with PowerPoint slides. But still, taking your own notes will help you better remember key points and concepts that are important to you, and quickly recollect relevant thoughts and ideas when you’re revising for the final round.
5. Final Preparation
After you’ve completed all the necessary topics and materials, this Exam Readiness: AWS Certified Machine Learning – Specialty (MLS-C01) course from AWS is a must-do. It’s a 4 hours free online e-learning course. You will learn valuable key test-taking strategies and skills in interpreting exam questions.
When faced with questions that you’re not confident with, you need to know how to tackle them. You will learn to identify important keywords, eliminate irrelevant or distractor options, and zoom down to the possible answer choices.
As most questions are scenario-based, you also need to learn to pay attention to the requirements specified in the question. For example, you may be asked to "select the most cost-effective solution" or "select a solution with minimal management and overhead".
The course will run through the 4 exam domains at a high level. For each domain, you will get to answer a few quiz questions.
You will also find an additional 35 study questions at the end. These questions mimic the type of questions that would appear on the actual exam. However, I still feel that the actual exam question’s difficulty level is higher. But I did get a few easy questions too.

For the final stint, try to take as many practice tests or practice questions as possible. You need to do this to get familiarized with the question style and be comfortable in answering them within the allocated time.
Do this even if you feel that you’re well prepared. Very often you may be surprised to discover that there are still one or two questions that you need to research further to find the answers to!
Again you will be able to find practice tests on platforms like Udemy or take the official practice exam from AWS which will cost 40 USD. If you have previously passed an AWS certification exam, you can take the AWS practice exam for free by utilizing the practice exam voucher from your AWS certification account.
Update: AWS had retired the benefit for a free practice exam. Starting on November 19, 2021, everyone can access AWS Certification Official Practice Question Sets on AWS Skill Builder for free.
Here are the links to some other free sample exam questions: Tutorials Dojo, ExamTopics, ML Exam, Whizlabs

Summary
In this post, I’ve briefly walked through the exam format and test domains for AWS Certified Machine Learning – Specialty.
I presented the various learning options that one can take before sitting for the exam. I also provided a summary of links to the diverse AWS resources and documentation that can help you in your exam preparation.
I recommend taking your own notes while going through learning materials. The notes will be very handy when you do your final rounds of revision.
For the final preparation, take the AWS exam readiness course. It is essential to learn test-taking strategies. Try out as many practice tests as possible.
Happy learning and good luck on your exam!
Before You Go…
🙏 Thank you for reading this post, and I hope you’ve enjoyed learning about the resources to prepare for the AWS Certified Machine Learning – Specialty exam.
💬 Please take a moment to share your thoughts and comments. Your feedback is important in supporting my work and helping to improve future content.
👉 If you like my post, don’t forget to hit Follow and Subscribe to get notified via email when I publish.
😃 Optionally, you may also sign up for a Medium membership to get full access to every story on Medium.
📑 Visit this GitHub repo for all codes and notebooks that I shared in my posts.
© 2024 All rights reserved.
Interested in learning about Transformers? I have the following posts for you.
Transformers, can you rate the complexity of reading passages?