How I passed the AWS Certified Machine Learning Specialty (MLS-C01) exam

My own personal study guide, and some resources and tips that will help you pass the exam!

June 12th, 2019

In March of 2019, AWS released the AWS Certified ML exam from beta. I had been eyeing this exam for some time, especially as a way to prove some of my prior ML experience, so I knew I wanted to study and pass this one.

After studying for about the last month, I was able to achieve a passing score on the exam on my first try! Now, this exam may seem simple for many people, or it may seem daunting for others. I hope to write my guide in a way for those who are more in the daunting side. This exam is absolutely doable, but it is not an easy exam by any means. This exam is on the same level as the other AWS professional certs. So, study hard, and hopefully I can offer a few tips to guide your study and build your confidence for the exam.

The basics of the exam

As of June 2019, the exam is around 65 questions with 180 minutes to do it. The majority of the questions are multiple choice, but there were also plenty of ‘choose 3 of 6’ etc. type of questions as well.

You are scored out of 1000. A passing score is a 750. No partial credit will be given on questions, so on those ‘choose X of Y’ questions, no credit is given if you only get some of them right. But you aren’t penalized for guessing, so if you don’t know it, then do your best and put down what you think.

There are 4 domains or topics that you will be tested on. Each domain is also weighted, as shown below:

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

For a full breakdown of the exam, then check out the exam blueprint in the resources section at the bottom.

My study plan

After taking the exam, and pondering about the questions from the exam, here is my breakdown of what helped prepare me for the exam and what gave me the most value for my time.

55% - A Cloud Guru
20% - Academic Statistics & ML concepts
15% - Solutions Architect Associate Cert
5% - White papers, FAQs, etc.
5% - AWS.Training

A Cloud Guru

At the time of writing this, there are very few options for the ML exam courses, because the exam only came out a few months ago! Luckily, the teams at A Cloud Guru set up a great course that I would absolutely recommend for anyone getting ready to study for this exam (I promise I have no affiliation with them but I wouldn’t mind them sending me a T-Shirt either 😁).

You can find there course here: AWS Certified Machine Learning – Specialty 2019 Course

I felt that the exam does a great job about introducing many of the areas of study that you will need to know. In addition, they also have quizzes that you can take to check your understanding, as well as hands on labs to gain some experience in services that you may have never used before. I took notes using Quizlet by just creating study cards as I watched the videos. This worked well and gave me lots to review when I had downtime as well.

I took fairly terse notes, and some of what I wrote is probably a little cryptic to anyone but me, but feel free to review my quizlet that I made to study: MLS-C01 Quizlet

I ended up signing up for 2 months of A Cloud Guru. I used the first 2 weeks to study for the Solutions Architect exam (which I’ll hit on in a bit) and the last monthish to go through the ML cert. I have about a half hour bus ride to work, so it was a great resource to watch to and from work, and then I would go through flash cards, the hands-on labs, or other resources at night.

If you decide against A Cloud Guru, then I would at least recommend going through some type of structured certification course so that you gain the breadth of the topic that you will need.

Academic Statistics & ML concepts

Now this is a difficult section because I largely drew on my own academic background from school where I studied topics like Statistics, Data Mining, Data Visualization, Running Models in R, etc.

So, what I would recommend is going through another course on Udemy or another platform that helps give you a basic, but solid, understanding of Machine Learning. I’m not saying you should need to be able to write out how to calculate models by hand with your math PhD. But rather, understand the concepts and how you can run many of the models using a library.

For example, play around with Pandas, Numpy, SciPy, and other basic Python libraries so that you can get familiar creating basic models. This will help you feel comfortable when you are creating Jupyter Notebooks in Sagemaker and should give you an understanding of some of the basic principles to ML. Kaggle is a great resource to find large datasets for you to practice on.

Solutions Architect Associate Cert

It is strongly recommended that for the ML Specialty cert, that you also obtain at least an associate level cert. So I decided to study for the Solutions Architect Associate Cert in order to help me prepare.

When I signed up for A Cloud Guru, I did both exams back to back, so that I would feel more prepared for the ML Specialty cert. I think this really helped me fill in many of the gaps for my understanding of AWS, so I would highly recommend this same path for anyone else, especially if you aren’t used to using AWS.

I have been using AWS for the past few years now, but studying for this exam grew my knowledge of a few areas further, and I think that there were dividends in the end when I was studying for the ML cert.

AWS Resources (White papers, FAQs, etc.)

Although it was a little boring at times, I did read (probably more like skimmed) through many of the white papers and FAQs of the different services. I didn’t feel like I gained a huge benefit from doing this, but I do think that it helped me with a few of the finer details of the services that you need to know.

So, I wouldn’t spend a ton of time reviewing all of these, but I would at least do a quick Passover of many of the more standard services. I’ll add links to these services and papers at the bottom of this post for easy access.

AWS.Training

AWS.Training was also a good resource. I think that it helped with growing a deeper understanding of many of the principles of ML. Some of the content is fairly high level, but I think that skimming through some of the content is still worthwhile. AWS.Training is actually the resources that AWS employees are given for learning ML. You have to pay for the labs, but the videos are all free. The UI has somewhat of a learning curve, but once you figure it out, then you can reap the free benefits.

Follow the Exam Prep path but if other videos interest you, then I would allow yourself to dive into those as well. Studying for an exam is hard, but it is much harder if you don’t enjoy what you are doing.

Final Advice

You can do it! This is a tough exam, but it will be that much more rewarding if you pass. I am happy to let this be a living document as well, so as you have questions, please reach out to me on Twitter or LinkedIn, and I will try to answer your questions by updating this post!

Good Luck!✌️

Recommended Resources

A Cloud Guru
Machine Learning Course - https://acloud.guru/learn/aws-certified-machine-learning-specialty
Solutions Architect Associate Course - https://aws.amazon.com/certification/certified-solutions-architect-associate/

AWS Exams
Exam Blueprint - https://d1.awsstatic.com/training-and-certification/docs-ml/AWS%20Certified%20Machine%20Learning%20-%20Specialty_Exam%20Guide%20(1).pdf
AWS Training - https://www.aws.training/Dashboard/?cta=tctopbanner
Learning Paths - https://aws.amazon.com/training/learning-paths/machine-learning/exam-preparation/

LinkedIn
Another summary of the cert - https://www.linkedin.com/pulse/aws-certified-machine-learning-specialty-some-thoughts-david-gu/

Udemy
Reinforcement Learning - https://www.udemy.com/artificial-intelligence-az/
Computer Vision with OpenCV - https://www.udemy.com/hands-on-computer-vision-with-opencv-python/

Blog Posts
Precision vs Recall (Confusion Matrix) - https://www.mikulskibartosz.name/precision-vs-recall-explanation/

AWS Misc. resources
Apache Spark - https://docs.aws.amazon.com/emr/latest/ReleaseGuide/emr-spark.html
Sagemaker built in Algorithms - https://docs.aws.amazon.com/sagemaker/latest/dg/algos.html

AWS Certs
Schedule and manage certs - https://www.aws.training/Certification