AWS Machine Learning Specialty (AWS)
AWS Certified Machine Learning - Specialty
More study resources
Cheat sheet
Keep the key ideas within reach.
ExploreKey terms
Get comfortable with the vocabulary.
ExploreStudy plan
Give each week a clear focus.
ExploreCommon mistakes
Understand the traps and how to avoid them.
ExploreCase studies
Put your knowledge to work in context.
ExploreExam-day guide
Prepare for the day ahead.
ExploreAbout this exam
The AWS Certified Machine Learning – Specialty credential is an advanced certification for practitioners who build, train, tune, deploy, and operate machine learning workloads on AWS. It suits data scientists, ML engineers, and cloud architects who already know the fundamentals of statistics, model evaluation, and at least one ML framework, and who now want to prove they can make those skills work inside the AWS ecosystem. The exam spans four areas: data engineering (ingesting, transforming, and storing data with services such as Glue, Kinesis, and S3), exploratory data analysis (feature engineering, visualization, and statistical analysis), modeling (algorithm selection, training, tuning, and evaluation), and ML implementation and operations (deployment, monitoring, security, and cost). Preparation should be hands-on rather than purely theoretical: build pipelines with SageMaker, train and tune models, deploy endpoints, and automate workflows. Supplement labs with AWS documentation, whitepapers on the Well-Architected Machine Learning Lens, and a review of core statistics and classic algorithms. This guide is independent study material and is not affiliated with or endorsed by Amazon Web Services; always confirm exam logistics and domain weightings against the official AWS exam guide before booking.
What you'll study
About AWS Machine Learning Specialty preparation
What is the format of the exam?
It is a multiple-choice and multiple-response test delivered at a testing center or online through remote proctoring. Some questions have one correct answer; others ask you to select two or more. The official AWS certification page publishes the current question count, time limit, and scoring details.
How hard is it, and what background do I need?
It is widely considered one of the tougher AWS certifications because it assumes real ML knowledge, not just cloud familiarity. Comfort with statistics, model evaluation metrics, and at least one framework such as scikit-learn or TensorFlow is essential. Without hands-on SageMaker experience, expect a long ramp-up.
How long should I study, and in what order?
Most candidates need several weeks of focused effort. Start with data engineering and exploratory analysis, since modeling questions assume you can prepare data. Then move to modeling and algorithm selection, and finish with deployment, monitoring, and security. Labs reinforce each stage better than reading alone.
Can I retake the exam if I fail?
Yes, AWS allows retakes, but there is typically a waiting period and a fee for each attempt. The exact waiting window and pricing change over time, so check the official AWS certification policies page for the current rules before scheduling a second attempt.
What are the most common failure modes?
Candidates often lose points by confusing similar services, such as when to use Kinesis versus Glue versus Batch, or by misreading metric trade-offs like precision versus recall. Others neglect operations topics such as monitoring, retraining triggers, and cost optimization because they focus only on modeling.
Do I need to memorize every algorithm?
No. You need to know which algorithm families fit which problem types, how they scale, and how to evaluate them. Understanding when to choose linear models, tree ensembles, or deep learning, and how SageMaker built-in algorithms differ, matters more than memorizing formulas.
Is hands-on experience required, or is reading enough?
Reading alone is risky. Many questions describe realistic scenarios where the answer depends on service behavior you only internalize by using it. Build at least one end-to-end pipeline: ingest data, engineer features, train and tune a model, deploy an endpoint, and set up monitoring.
How much does the exam cost and how long is it valid?
Fees and validity periods are set by AWS and change periodically. Rather than relying on secondhand figures, consult the official AWS certification page for the current price, recertification requirements, and validity duration before you register.
What score do I need to pass?
AWS uses a scaled scoring system, and the passing threshold is published by the official certification body rather than fixed in advance for candidates. Check the AWS certification page for current scoring information; do not rely on pass rates quoted in forums.