The AWS Certified Machine Learning - Specialty exam (MLS-C01) is retired: March 31, 2026 was its last testing day. Existing certifications remain active for three years from the date earned. Use this guide to understand your badge’s status, revisit useful modeling concepts, and plan study for a separate current credential.
Can you still take the Specialty exam?
No. March 31, 2026 was the last day to take MLS-C01, including retakes for recertification.
AWS’s dated retirement notice and retirement FAQ establish the final testing date. Generic scheduling or renewal text lower on an older page does not reopen the retired exam.
This article keeps the MLS-C01 identity and its historical syllabus clear. The separate Machine Learning Engineer - Associate credential has its own scope, exam version, and requirements.
Sources: AWS certification pages, the official MLS-C01 exam guide, the Machine Learning Engineer - Associate page, and AWS machine learning documentation, retrieved September 14, 2026.: AWS Certified Machine Learning - Specialty (AWS certification page) — Retirement notice, exam overview, and retirement FAQs
What happens to a badge you already earned?
Your Specialty certification remains active for three years from its earning date. AWS says holders need take no action because of retirement.
You may continue displaying and sharing the digital badge through Credly during that validity period. Check your own certification record for its expiry date.
Retaking MLS-C01 is no longer a renewal option. A separate certification does not automatically rename or convert your existing Specialty badge.
Sources: AWS certification pages, the official MLS-C01 exam guide, the Machine Learning Engineer - Associate page, and AWS machine learning documentation, retrieved September 14, 2026.: AWS Certified Machine Learning - Specialty (AWS certification page) — Retirement notice, exam overview, and retirement FAQs
What the old syllabus can still help you review
The historical MLS-C01 syllabus covered data engineering, exploratory data analysis, modeling, and ML implementation and operations.
Use its topic map to organize existing notes and identify useful concepts to revisit. The percentages below belonged to the retired Specialty exam and do not describe the current Associate exam.
When preparing for another credential, map each lesson to that exam’s current tasks and verify service behavior in current documentation. The historical guide’s tasks and service lists were not exhaustive.
| Content domain | Share of scored content | What it covered |
|---|---|---|
| Data Engineering | 20% | Data repositories for ML, ingestion solutions for batch and streaming workloads, and transformation pipelines |
| Exploratory Data Analysis | 24% | Sanitizing and preparing data, feature engineering, and analyzing and visualizing data |
| Modeling | 36% | Framing ML problems, model selection, training, hyperparameter optimization, and model evaluation |
| Machine Learning Implementation and Operations | 20% | Scalable and resilient ML solutions, service selection, security practices, and deployment and operationalization |
Sources: AWS certification pages, the official MLS-C01 exam guide, the Machine Learning Engineer - Associate page, and AWS machine learning documentation, retrieved September 14, 2026.: AWS Certified Machine Learning - Specialty Exam Guide (MLS-C01) — 2026 edition; content outline, scoring, and scope lists
Practice: look beyond overall model accuracy
Accuracy, precision, and recall answer different questions about a binary classifier. Use all three to understand the supplied results.
A true positive is an actual positive predicted as positive; a false positive is an actual negative predicted as positive. A false negative is an actual positive predicted as negative, and a true negative is an actual negative predicted as negative.
Accuracy is the fraction of all predictions that are correct. Precision is the fraction of predicted positives that are actually positive. Recall is the fraction of actual positives identified as positive. Which measures matter most depends on the application’s goals and error costs.
The reference for these definitions is archived Amazon Machine Learning documentation. This exercise uses the stable metric definitions; it does not require that legacy service or its console.
Practice: the metric drill is an original study exercise, not a real exam item.
An original study example evaluates a fraud-screening classifier on 1,000 validation cases at one fixed threshold. The results are 40 true positives, 10 false positives, 20 false negatives, and 930 true negatives. Calculate accuracy, precision, and recall. Which metric shows the proportion of actual fraudulent cases detected, and why is 97% accuracy alone an incomplete description?
Work through the calculations before reading the answer.
Accuracy is (40 + 930) / 1,000 = 97%. Precision is 40 / (40 + 10) = 80%. Recall is 40 / (40 + 20) ≈ 66.7%. Recall measures the proportion of actual positives detected.
There are 60 actual fraudulent cases: 40 detected and 20 missed. Overall accuracy counts 930 correct negative predictions alongside the 40 correct positives, so 97% accuracy can coexist with missing one-third of the positive cases. Precision answers a different question: 40 of the 50 alerts are true positives. These figures describe the supplied validation results at one threshold; they do not establish the right operating threshold for every application.
Sources: AWS certification pages, the official MLS-C01 exam guide, the Machine Learning Engineer - Associate page, and AWS machine learning documentation, retrieved September 14, 2026.: Binary Model Insights (Amazon Machine Learning developer guide) — Confusion matrix outcomes, precision, recall, accuracy, and score cut-off trade-offs
Considering Machine Learning Engineer - Associate
AWS points machine learning candidates to the separate Machine Learning Engineer - Associate credential, focused on implementing and operating ML workloads in production.
Check the official page for the version and language you intend to take. As of September 14, 2026, the English MLA-C01 exam remains available through September 28, and English beta delivery for the updated MLA-C02 begins September 29.
Beta registration opened on September 1; that was not the first testing date. The Japanese, Korean, and Simplified Chinese MLA-C01 exams remain available until MLA-C02 general availability, whose date the page lists as to be determined.
Use the guide for your selected version. The old Specialty’s domain weights and preparation resources do not establish the current Associate syllabus.
Sources: AWS certification pages, the official MLS-C01 exam guide, the Machine Learning Engineer - Associate page, and AWS machine learning documentation, retrieved September 14, 2026.: AWS Certified Machine Learning Engineer - Associate (AWS certification page) — MLA-C02 beta notice, exam overview, and key dates
A useful plan for your next step
Keep the skills that remain relevant, and match any new exam preparation to its current guide.
The following are study suggestions. Choose a credential or learning path that matches the work you want to do, then use its published tasks to identify gaps.
- If you hold MLS-C01, record your existing badge’s expiry date.
- For a new certification, confirm the exam version, language, and testing date before choosing materials.
- Map useful older notes to the current objectives, marking material that needs updating.
- Practice explaining model-evaluation results, including which errors a metric exposes or hides.
- Use current technical documentation and relevant practice resources for your chosen path.
- Revisit gaps found in practice rather than assuming the old Specialty syllabus transfers unchanged.
Sources: AWS certification pages, the official MLS-C01 exam guide, the Machine Learning Engineer - Associate page, and AWS machine learning documentation, retrieved September 14, 2026.: AWS Certified Machine Learning - Specialty (AWS certification page) — Retirement notice, exam overview, and retirement FAQs; AWS Certified Machine Learning - Specialty Exam Guide (MLS-C01) — 2026 edition; content outline, scoring, and scope lists; AWS Certified Machine Learning Engineer - Associate (AWS certification page) — MLA-C02 beta notice, exam overview, and key dates; Binary Model Insights (Amazon Machine Learning developer guide) — Confusion matrix outcomes, precision, recall, accuracy, and score cut-off trade-offs
The historical exam format
Before retirement, MLS-C01 used 65 questions in 180 minutes. These details are historical reference, not current booking information.
The exam had 50 scored questions and 15 unidentified unscored questions, using multiple-choice and multiple-response formats. It required an overall scaled score of 750 on a scale of 100–1,000, rather than a separate pass in each domain. A scaled score is not a raw percentage correct.
The listed fee was USD 300 before applicable taxes and currency adjustments. The guide targeted candidates with two or more years of relevant AWS machine learning experience; that background was recommended preparation.
Sources: AWS certification pages, the official MLS-C01 exam guide, the Machine Learning Engineer - Associate page, and AWS machine learning documentation, retrieved September 14, 2026.: AWS Certified Machine Learning - Specialty (AWS certification page) — Retirement notice, exam overview, and retirement FAQs; AWS Certified Machine Learning - Specialty Exam Guide (MLS-C01) — 2026 edition; content outline, scoring, and scope lists
Sources: AWS certification pages, the official MLS-C01 exam guide, the Machine Learning Engineer - Associate page, and AWS machine learning documentation, retrieved September 14, 2026.
Facts checked against retrieved AWS sources on September 14, 2026; retirement dates follow AWS's dated notice.: