Directory / Amazon Web Services

AWS Certified Machine Learning – Specialty

Formerly the deepest AWS machine learning exam, with over a third of the blueprint on modelling. It contains no generative AI content.

What it signals

Practical assessment: Scenario-based multiple choice. Proctored scenario-based multiple choice/multiple response.
Identity assurance: Proctored with identity verification. Pearson VUE testing centre or online proctored (while available)
Depth: Specialty or expert. Specialty level.
Experience expected: 2 to 3 years. Guide recommends 2+ years.
Portability: Mostly transferable. Large share is general ML theory (algorithms, metrics, feature engineering) alongside SageMaker specifics.
Verifiability: Public badge or registry. Credly digital badge with verification.

For hiring managers

Retired: nobody can earn it after March 31, 2026, but it remains a valid and respected signal on a CV until the holder's 3-year expiry.
It was the deepest AWS classical ML exam (36% modelling). No generative AI content.
AWS points to AI Practitioner, ML Engineer - Associate and Data Engineer - Associate as alternatives; none is a like-for-like replacement for the modelling depth.
Check the expiry date on the Credly badge.

Verifying it

Credly digital badge link shared by the candidate; badge is active only while the certification is valid.

What it covers

Published exam outline

Data Engineering 20%
Exploratory Data Analysis 24%
Modeling 36%
Machine Learning Implementation and Operations 20%

By topic

GenAI apps: Not covered. Blueprint predates foundation-model application development.
Agents: Not covered. No agent content.
Classical ML: Core focus (over 35%). Modeling is 36% and much of Exploratory Data Analysis (24%) supports model building.
Operations: Substantial (20 to 35%). ML Implementation and Operations is 20%.
Data: Substantial (20 to 35%). Data Engineering is 20%, plus data preparation within EDA.
Responsible AI: Touched on (under 10%). Security appears only as a task within Domain 4.

Exam details

Who the provider says it is for. Individuals in an AI/ML development or data science role who design, build, deploy, optimise, train, tune and maintain ML solutions on AWS.

Format: multiple choice; multiple response
Delivery: Pearson VUE testing centre or online proctored (while available)
Code or hands-on work: No labs and no code writing; multiple choice and multiple response only.
Prerequisites: None
Recommended experience: 2 or more years of experience developing, architecting, and running ML or deep learning workloads in the AWS Cloud.

Preparing

Not confirmed

These details were not available from an official source at the time of research.

Whether retired-certification holders can extend via Skill Builder maintenance
Exact date the retirement was announced (blog is October 2025; day not captured)

Sources (4)