AWS has deeply reworked its artificial intelligence certifications. The Machine Learning – Specialty certification was retired on March 31, 2026, and a more progressive path replaces it, aligned with how teams actually work today: less training of in-house models, many more applications built on foundation models.

Why the change

Most companies no longer train their own models from scratch. They use managed foundation models (for example through Amazon Bedrock) and integrate them into their applications, with agents, knowledge bases and guardrails. The certifications follow this shift: the focus moves from machine learning theory to putting generative AI applications into production.

The new AWS AI path

Certification Code Level For whom
AWS Certified AI Practitioner AIF-C01 Foundational Business or technical beginners, up to 6 months of exposure to AI on AWS
AWS Certified Machine Learning Engineer – Associate MLA-C01 Associate Engineers who build, deploy and maintain ML pipelines (MLOps)
AWS Certified Generative AI Developer – Professional AIP-C01 Professional Developers and AI engineers who ship solutions built on foundation models

AI Practitioner (AIF-C01): the entry point

Five domains: fundamentals of AI and ML (20%), fundamentals of generative AI (24%), applications of foundation models (28%), responsible AI (14%), and security, compliance and governance (14%). It is the right door if you are new to the topic.

Machine Learning Engineer – Associate (MLA-C01): MLOps

Four domains: data preparation (28%), model development (26%), deployment and orchestration (22%), and monitoring, maintenance and security (24%). It covers most of the hands-on skills of the former Specialty.

Generative AI Developer – Professional (AIP-C01): the newcomer

Out of beta since March 31, 2026, it targets people with about two years of cloud experience and at least one year of hands-on generative AI work. It was refreshed to include Amazon Bedrock AgentCore, a sign that agents are now at the heart of real-world use. AWS also offers a hands-on agentic AI microcredential, assessed in a live AWS environment.

What if you hold the Machine Learning – Specialty?

It remains valid until its expiration date. Going forward, MLA-C01 covers the MLOps side and AIP-C01 the generative AI side: choose based on your role.

Worth noting: the Security – Specialty certification also moves to a new version (SCS-C03), with more generative AI and ML security. AI is becoming a cross-cutting skill, security included.

Which order to follow?

  1. Beginner: AI Practitioner (AIF-C01).
  2. Data / MLOps profile: Machine Learning Engineer – Associate (MLA-C01).
  3. AI application developer: Generative AI Developer – Professional (AIP-C01), ideally after an Associate certification (Developer, Solutions Architect or ML Engineer).

Whatever the level, the method stays the same: start from the official exam guide, practice in an AWS account and train with timed mock exams. AWS courses are coming to the OnLevelUp catalog.

Sources