Advice & Tips for Passing AWS Machine Learning Specialty

Reflections & advice on passing the CLF-C01 & MLS-C01 in late 2021

Published in

Towards Data Science

7 min readJan 30, 2022

In the second half of 2021 (aka. day 673 of the year 2020), I took and passed the AWS Certified Machine Learning Specialty exam. I read a bunch of other articles on the subject, all of which helped and supported me on the path to passing the exam. After finishing I wanted to contribute to that pool of knowledge so, in true data science fashion, I thought I’d go ahead and open source my experience and preparatory materials.

Background

Everyone approaches these exams differently, so before I get into the meat of the article, it may help to detail my background and where I started studying for the exam. At the time I began studying, I had been practicing data science for approximately 7 years. I held a masters in applied statistics and did a minor in math during my undergraduate years. At the point of taking the exam, the extent of my experience with cloud-based computing consisted of having trained and set up a deployment of XGBoost on SageMaker. Despite being greener than the tree-based algorithm I’d trained, I set my sights on light, fluffy cloud greatness and got going.

Exams Timeline and Length of Preparatory Time

Now that we’ve gotten the preliminary info out of the way, let’s get into what the path to the exam looked like for me.

As many before have advised, I chose to take the Cloud Practitioner first (Adam DeJans does a great job at explaining why you would want to do this (both from a preparatory standpoint and a financial one). I signed up for the practitioner exam at the end of September and started studying in earnest after Labor Day weekend. Once I dug in, I found that the Cloud Practitioner exam, while broad in scope, was not as difficult of an exam — roughly three weeks was plenty of time to get a handle on the material.

After passing the Cloud Practitioner exam, I took a weekend off before starting to study for the mid-November Machine Learning Specialty exam. This test required a lot more of my attention, and I began studying in early October to give myself a full 6 weeks to prepare. I work full-time so to preserve my sanity and get the occasional glimpse of the outside world, I chose to schedule my study sessions for 2 hours roughly 2–3 nights a week with longer study sessions on the weekend. All in all, I gave myself 3 weeks for the Cloud Practitioner exam and 6 weeks for the Machine Learning Specialty exam, which I found was sufficient time to cover the material thoroughly while working a full-time job.

The main strategy I employed during my study times was going through Udemy courses, then spending the rest of my time taking practice exams and reviewing the source material for answers I missed. The practice exams helped identify gaps in my knowledge and helped me prepare for the testing of that knowledge on the exam (for example, distinguishing between similar services such as Lex/Transcribe/Polly/Comprehend to be pretty important). One final note: I did not go over white papers (I personally prefer pink papers with little doodles in the corners) or use the formal AWS Trainings, though I did watch a few of the Emily Webber and previous re:Invent conference videos.

Since the MLS exam was more difficult and fewer people have shared about their experience of this exam, I’ll spend the rest of the post focusing more specifically on the MLS exam (though I’ll also reference some articles and resources below that were helpful in passing the CPP).

Exam Experience

Since I signed up for both exams in the midst of the pandemic, I opted to take them from home with Pearson OnVue instead of going to a testing center (I’d risk getting COVID for In-N-Out’s animal fries, but not the CPP). Obviously I can’t reveal what was on the exam without an AWS hit squad busting through my window, but I’ll share a few thoughts on what the exam experience was like for me.

Like many others who are approaching the exam with a similar math/stats background, the easiest parts of the exam for me were the data analytics and modeling portions. Examples included deciphering charts, interpreting a confusion matrix (the name totally checked out), tuning specific algorithms like random forests, and more. I personally had more difficulty with parts of the exam covering specific use cases and configurations of AWS services (determining appropriate parameters for specific AWS services, specific order of folder structures, etc). Thanks to the CPP I was able to readily identify AWS services, but because the MLS-C01 is a professional-level exam, it goes much deeper into the implementation of these services than I felt prepared for. I ended up taking the full 3 hours because I had flagged questions I was uncertain about and spent time at the end backtracking to those. At the end of the exam, I wasn’t left in suspense — the testing platform brought me through a short survey, then gave instant feedback on whether I passed. I must have still been riding on pure adrenaline, because I sweat bullets all the way through the ~10 question survey.

Study Materials & Resources

What was helpful in other blog posts and articles were the list of materials and resources each writer found most useful. I’ve compiled my own list, so here are the materials that helped me the most throughout the process:

First, I really appreciated this Reddit post. The author shares some great analysis on his practice exam results prior to taking the exam and covers the resources he used in detail.
These are the courses I went through on Udemy —

Recommend strongly
AWS Certified Machine Learning Specialty 2022 — Hands On! | Udemy

Good, but less helpful than the above
AWS Machine Learning Certification Exam | Complete Guide | Udemy
Amazon AWS SageMaker, AI and Machine Learning with Python Course | Udemy

These practice exams (also on Udemy):
AWS Certified Machine Learning Specialty Practice Exams | Udemy
AWS Certified Machine Learning Specialty: 3 PRACTICE EXAMS | Udemy
AWS Certified Machine Learning Specialty Full Practice Exam | Udemy

Quizlet. Between August and November 2021, Quizlet replaced Instagram and became my best friend. I used Quizlet throughout my Udemy courses as a repository for unfamiliar terms and info, as well as a repository of questions I’d gotten wrong and wanted to revisit. Here are my Quizlet decks and other decks I found helpful:

Final Pieces of Advice

Looking back on my personal process, here are some advice I would have given my past self:

Having more SageMaker experience would have made the exam experience easier
The exam creators recommend having at least 2 years of experience in using AWS’ ML/AI services and platforms. I probably would not have struggled as much if I had waited to take the exam and had worked more with SageMaker
The exam is not about accuracy — it is about knowing more than the curve.
You will be judged against those who have gotten the question right in the past, not on overall accuracy. This article explains the scoring criteria Demystifying your AWS Certification exam score | AWS Training and Certification Blog
The weekend before the exam, I was retaking exams and averaging about 70% accuracy, which made me uneasy because the practice exams estimate 75% accuracy as the benchmark for passing
Schedule the exam at a part of the day where you would be the most alert
I initially took the CPP from 3pm — 5pm and I could feel my brain slowing down so for the Machine Learning Specialty exam I scheduled it for earlier in the day (11:30am — 2:30pm) and felt a lot more alert
Fuel up well before the exam
There are a lot of studies and nutritional experts that talk about the correlation between food and our brain’s performance. I knew from my own personal experience that when my body is in a fasted state, my mind tends to be clearer. So I restricted calories the week before and the day of the exam. And for me, my brain was at its optimal state and I experienced less mental fog during the exam.
Reading the questions carefully and implement practices to slow down
Sometimes I can read the question to fast and miss key details so I used the note-taking functionality on the Pearson platform and typed out key words and phrases from the question to slow myself down to focus on important details of the question
Use the question flagging feature
I used the flagging feature on practice exams and the actual exam to distinguish the difficult questions from the easy ones so when I went back to review my responses I could skip over the easy questions
Redeem the free practice test voucher right before you are ready to take the practice exam
Right after you pass any certification, you can redeem a voucher for a free practice test but what I didn’t realize is that the practice test expires a month after you redeem it — so my recommendation is to redeem it close to the time that you plan on taking it — I mistakenly thought I would have unlimited access to the timed exam
It is permissible to have water on your at-home exam table
Finally with Pearson, you can have a glass of water on the exam table but it isn’t advised to use the restroom so I had to sit it out for the entire 3 hours. Also don’t forget to have your laptop plugged in and your cell phone nearby (but not on the table)

I tried my best to capture my experience of studying, preparing and taking the exam — hopefully this helps you on your journey! I would be happy to answer any further questions on Twitter(or find me on LinkedIn) and connect with you on the other side of your test! Thanks for reading — good luck, and hope to see you at future AWS events!

Cloud Practitioner Resources

AWS Certified Cloud Practitioner Exam Training [New] 2022 | Udemy
[Quizlet Deck] AWS Cloud Practitioner Services
[Quizlet Deck] AWS CPP Quiz Questions