AI and ML Certification Guide 2026: Two of the Big Three Exams Retired — What Replaced Them
If you are following an AI/ML certification roadmap written before 2026, two of the three exams on it no longer exist. The AWS Certified Machine Learning – Specialty exam had its last testing day on March 31, 2026. Microsoft’s Azure AI Engineer Associate certification, the AI-102 path, is retired along with its renewal assessment. Only the Google Cloud Professional Machine Learning Engineer certification survived the year intact, and even that one has had its content rebuilt around the Gemini Enterprise Agent Platform. This page is a platform-choice guide first and an exam-prep guide second: what is actually live as of August 2026, what each live exam costs and how long it runs, and which direction fits the stack you already work in.
As of August 2026 the three live vendor anchors are Google Cloud Professional Machine Learning Engineer (120 minutes, 50–60 questions, $200), AWS Certified Machine Learning Engineer – Associate (130 minutes, 65 questions, $150), and Microsoft’s replacement pair, Exam AI-103 for Azure AI Apps and Agents Developer Associate and Exam AI-300 for Machine Learning Operations Engineer Associate, both 120 minutes. AWS ML – Specialty and Azure AI-102 are retired.
Last updated on August 30, 2026
How we verified this page
Every certification on this page was checked against the vendor’s own certification page on August 30, 2026: Google Cloud, AWS, and Microsoft Learn. The AWS ML – Specialty retirement date and the Azure AI Engineer Associate retirement notice are quoted from those official pages. Microsoft’s AI-103 page was last updated 2026-08-11 and the AI-300 page 2026-08-13. Wage data comes from the U.S. Bureau of Labor Statistics Occupational Outlook Handbook, May 2025 figures. Vendor exam pricing varies by country and is subject to change — confirm on the vendor page before you register.
What is live, what is retired, and what each exam costs
| Fact | Current detail | Source |
|---|---|---|
| RETIRED — AWS Machine Learning – Specialty | The last day to take the AWS Certified Machine Learning – Specialty exam was March 31, 2026. Existing holders keep active status for 3 years from their earning date | AWS ML Specialty certification page |
| RETIRED — Azure AI Engineer Associate (AI-102) | Microsoft states that this certification and its renewal assessment are retired. Any roadmap still pointing at AI-102 is out of date | Microsoft Azure AI Engineer Associate page |
| LIVE — Google Cloud Professional ML Engineer | 120 minutes, 50–60 multiple choice and multiple select questions, $200 registration, offered in English and Japanese | Google Cloud Professional ML Engineer |
| Google recommended experience | 3+ years of industry experience including 1+ years designing and managing solutions on Google Cloud. There is no formal prerequisite, but the exam expects you to read Python and SQL | Google Cloud Professional ML Engineer |
| Google content change | The exam was updated to reflect the move from Vertex AI to the Gemini Enterprise Agent Platform and changes across the Google Cloud data and analytics stack. Study material written before that update describes services the exam no longer centers on | Google Cloud Professional ML Engineer |
| LIVE — AWS ML Engineer – Associate | MLA-C01: 130 minutes, 65 questions, $150. This is the replacement path for the retired ML – Specialty exam | AWS ML Engineer Associate page |
| AWS version change | An updated version, MLA-C02, opens for registration on September 1, 2026. If you are booking now, check which version your appointment covers before you buy study material | AWS ML Engineer Associate page |
| AWS recommended experience | At least 1 year using Amazon SageMaker and other ML engineering AWS services, plus 1 year of hands-on experience with AWS services | AWS ML Engineer Associate page |
| LIVE — Microsoft Exam AI-103 | Azure AI Apps and Agents Developer Associate. 120 minutes. Covers planning and managing an Azure AI solution, generative AI and agentic solutions, computer vision, text analysis, and information extraction, built around Python and Microsoft Foundry | Microsoft Azure AI Apps and Agents Developer Associate |
| LIVE — Microsoft Exam AI-300 | Machine Learning Operations Engineer Associate. 120 minutes, offered in English only. Covers MLOps and GenAIOps infrastructure, model lifecycle and operations, generative AI quality assurance and observability, and system optimization | Microsoft Machine Learning Operations Engineer Associate |
| How to choose between the two Microsoft paths | AI-103 is the application and agent building path; AI-300 is the operations path for training, deploying, monitoring, and optimizing models. AI-300 is the closer analogue to the Google and AWS ML engineering exams | Microsoft Learn certification pages |
| Cluster reality | This is a platform-choice problem first and an exam-prep problem second. The cheapest and shortest live path is AWS at $150 and 130 minutes; the most expensive is Google at $200 | Google, AWS, and Microsoft official certification pages |
AI and ML salary benchmarks
Because this page spans multiple vendor paths, the salary section uses broad BLS occupation data rather than one provider-specific marketing claim.
| Role / benchmark | Pay | Why it matters |
|---|---|---|
| Data scientists | $120,230 median annual wage | May 2025 figure. Strong broad benchmark for AI and ML career paths BLS Data Scientists |
| Data scientist job outlook | 35% projected growth, 2025 to 2035 | Much faster than average, with about 24,800 openings projected each year BLS Data Scientists |
| Computer and information technology occupations | $109,470 median annual wage | May 2025 figure, against $50,980 for all occupations. Broad benchmark for technical careers where ML certification can become a differentiator BLS Computer and Information Technology Occupations |
Related jobs for this exam path
Use the exam-specific jobs page to see the role families, live search links, and hiring context tied to this certification or study path.
Start here
1. Run the quiz first
Use the free quiz to see whether you need fundamentals, project work, or communication practice next.
Open →2. Read the best-fit guide
Move into the article that matches your biggest bottleneck instead of reading the whole cluster randomly.
Open →3. Use the AI tutor
Get targeted explanations, follow-up questions, and a tighter next-step study plan.
Open →How to use these guides
Free quiz
Start with the quiz while your baseline is still honest. That gives the rest of this page a real job.
Open →Best guide article
Use the strongest article first, then expand only into the narrower guides that solve your specific weakness.
Open →AI tutor
The tutor is the explanation layer. It turns missed questions and vague uncertainty into a concrete next-step plan.
Open →Jobs page
See the related role families, live search links, and hiring intent connected to this exam cluster.
Open →Complete article library
Every article below belongs to the AI & ML Certification Prep cluster. The grouping is by funnel stage so the page works both as a study hub and as a cleaner internal-link destination.
Awareness
Start here if you are choosing between the AWS and Google ecosystems at entry level before committing to an engineer-tier exam.
Compares two paths directly so you can choose the stronger option before you commit more study time.
Consideration
Focused strategy piece for candidates who want a practical path to a pass or score target instead of broad background reading.
Useful supporting guide for this exam cluster when you need a deeper explanation than the main pillar can provide.
Decision
Breaks down the exam content itself so you can see what the test is really rewarding and where to focus first.
High-intent
Useful supporting guide for this exam cluster when you need a deeper explanation than the main pillar can provide.
Useful supporting guide for this exam cluster when you need a deeper explanation than the main pillar can provide.
Official resources
Official page for the only one of the original three anchor exams still running.
The replacement for the retired ML – Specialty exam, with MLA-C02 opening September 1, 2026.
States the March 31, 2026 final testing date and what happens to existing credentials.
Azure AI Apps and Agents Developer Associate — the app and agent building path.
Machine Learning Operations Engineer Associate — the MLOps and GenAIOps path.
External resources and benchmarks
Grounds the AI/ML value proposition in neutral salary and growth data.
Useful when you need to compare the ML engineer path with adjacent Google Cloud certs.
Next move
Use the quiz, the guide, and the tutor as one study loop
These pillar pages are meant to reduce scattered browsing. Start with the quiz, move into the article that matches your weakness, then use the tutor to get targeted feedback until the mistake pattern changes.
Frequently asked questions
Is the AWS Machine Learning – Specialty exam still available?
No. AWS states that the last day to take the AWS Certified Machine Learning – Specialty exam was March 31, 2026. People who already hold it keep active status for three years from the date they earned it. The replacement path is AWS Certified Machine Learning Engineer – Associate.
Should I still study for AI-102?
No. Microsoft states that the Azure AI Engineer Associate certification and its renewal assessment are retired. The current Microsoft paths are Exam AI-103 for Azure AI Apps and Agents Developer Associate and Exam AI-300 for Machine Learning Operations Engineer Associate.
Which live AI/ML certification is the cheapest and shortest?
AWS Certified Machine Learning Engineer – Associate, at $150 for a 130-minute, 65-question exam. Google Cloud Professional Machine Learning Engineer is $200 for 120 minutes and 50–60 questions. Microsoft prices AI-103 and AI-300 by country, and both run 120 minutes.
Is this page for one exam or several?
Several. This guide covers the live Google, AWS, and Microsoft AI/ML certification paths, because most candidates are choosing between platforms before they commit to a single exam.
Which path is best for beginners?
Usually none of these. Every live exam here assumes about a year of hands-on platform experience — AWS asks for a year with SageMaker, Google recommends three years of industry experience. This guide works best for people who already have a technical base and need to choose a vendor direction.