I’ve just earned all three AI/ML AWS certifications and want to share my subjective thoughts about each certificate in particular.
Inspiration
I decided to pursue these three AWS AI/ML certifications after attending re:Invent 2024 in Las Vegas. Nearly every talk revolved around AI - how to integrate AI into workloads or how AI is already transforming solutions both within and beyond AWS.
In my daily work as an engineer, I have zero AI/ML-related projects, which made me realize a significant gap in my knowledge. I felt the need to close the gap and expand my knowledge in this field.
Another motivation was my existing eight AWS certifications - I wanted to complete the full set.
AWS Certified AI Practitioner
The first certification I tackled was the AWS Certified AI Practitioner. For preparation, I followed AWS’s 4-step plan, available on Skill Builder.
To be honest, AI/ML concepts are challenging for me. There’s a huge amount of new terminology to learn and retain. One of the most frustrating aspects of the exam was the focus on specific terms.
For example, you might be asked to evaluate a model based on given requirements and choose the correct metric - Recall, Specificity, Precision, F1 Score, or Accuracy. Understanding these, along with the confusion matrix, is crucial.
Beyond that, the exam covers topics like:
- Monitoring AI workloads
- Security considerations
- Authentication & Authorization
- Data storage options for AI/ML
Since I already had the AWS Certified Solutions Architect - Professional certification, I focused solely on AI/ML content. The free AWS Skill Builder course is enough to pass, but be prepared — the instructor speaks very slowly. I ended up watching everything at 2x speed or even reading the transcript instead.
AWS Certified Machine Learning Engineer – Associate
The second certification I pursued was the AWS Certified Machine Learning Engineer – Associate. Following my previous experience, I decided to stick with AWS’s 4-step plan on Skill Builder.
I genuinely tried to follow the guide, but every time I started watching the videos, I found myself falling asleep within 15 minutes. The content itself is solid, but I believe it’s more about my personal learning style — I struggle to absorb information this way.
Looking for an alternative approach, I turned to my friend Dmytro Sirant for advice. He recommended the course AWS Certification Machine Learning Specialty 2025 by Stephane Marek and Frank Kane. Their explanations were clear, detailed, and engaging, covering all the key topics needed for the exam.
Interestingly, you could even use this course to prepare for the AWS Certified AI Practitioner exam. In my opinion, both certifications are quite similar in terms of difficulty and required knowledge. The main difference is that the Machine Learning Engineer – Associate exam features more wordy questions and demands a broader understanding of AWS domains, including databases, networking, and general cloud concepts. However, when it comes to AI/ML-specific content, the level of difficulty is comparable to the Practitioner exam.
AWS Certified Machine Learning – Specialty
The third and final certification I pursued was AWS Certified Machine Learning – Specialty. For preparation, I chose the AWS Certified Machine Learning Specialty Video Course on Digital Cloud Training, taught by Karim Elkobrossy and Neal Davis. This course provided a great refresher on AI/ML topics, with some concepts explained more clearly than in other courses I’ve taken. Overall, I found the platform very effective.
The exam itself is little a bit outdated — there were no questions on Bedrock, for example — but it remains hallenging. Surprisingly, I scored 50 points higher than on the associate-level exam, likely thanks to the additional two weeks of studying.
In terms of topics and difficulty, the Machine Learning – Specialty exam is quite similar to Machine Learning Engineer – Associate, so if you’ve prepared for one, transitioning to the other isn’t too difficult.
One key difference is exam duration:
- Associate-level exam: 130 minutes
- Specialty-level exam: 180 minutes
That extra 50 minutes for almost the same level of exam.
Final Thoughts
Despite being labeled a “practitioner level” certification, AWS Certified AI Practitioner felt more like an associate level challenge. The depth of AI/ML knowledge required was higher than I expected. You are expected to know a lot of terminology and, more importantly, understand how it works. While the exam does include some general AWS topics like security, IAM, and storage, the AI/ML questions remain quite difficult.
Both AWS Certified Machine Learning Engineer – Associate and AWS Certified Machine Learning – Specialty are at a similar difficulty level, and I would personally classify them both as specialty level exams. The associate level exam is newer and includes questions about Bedrock and other modern AI services. However, in order to pass either exam, you need strong foundational knowledge in AI/ML domains. This includes:
- Data preparation
- Data storage and management
- Model training and evaluation
- Model deployment and monitoring
My Advice
If you’re planning to take these certifications, I recommend:
- Starting with the Practitioner level certification to build foundational knowledge.
- Choosing either the Associate or Specialty level as the next step — whichever aligns better with your goals.
- Avoiding long gaps between exams. The content is similar across all three, so taking exams close together can help retain knowledge. I took mine with two-week breaks, using that time for learning and preparation.
If you’re considering the AWS AI/ML certification path, I hope my experience helps! Best of luck with your studies and exams! 🚀