I decided to take a new exam this year, to learn more about machine learning and data science in general. For me, one great path was related to the AWS Machine Learning Specialist. I don't have a huge background in those areas but I have been working with data for a few years now (and with two AWS certifications, one I explained here about AWS Data Analytics Specialist), I decided that was the time to take that exam and get more familiar with those fields.

After almost four months of studying one to two hours per day, I was able to get my certification. I intend to share a bit about my process, this might help someone out there that wants to know more about the certification and how to prepare. The steps that I used to organize my study:

  • Methods: which tools I used to study, what resources, and for how long each resource. Which materials do I recommend or suggest staying away from, and so forth.
  • Analysis: Take a look at the hours that were used to study and review questions.
  • Result: My experience with the exam, the score I received, and some learnings over that process. Plus, compare my score on that exam with other AWS exams that I did.

Methods

The best way that I felt to start was similar to my previous study for other AWS certifications:

  1. Start with a video course or a book based on the AWS content, usually to get a general idea of the services that I am not familiar with.
  2. Get as many relevant sample/practice questions as possible.
  3. Revisit and train with those questions. Study a specific topic if I fail frequently on a subset of questions. Usually with an SRS (Spaced Repetition System) system.

The main materials and tools I used for each step:

  1. Courses/Books: Used the following resources to study in that phase:
  2. Sample Questions: I used the following sources to collect practice questions and review them later on.
  3. Practice: I used mostly Anki here, with the questions that I collected in the previous step. I would take a print screen for each question that I collected and put the answer on the back of the card. While I was revisiting and studying, Anki would show more frequently the cards that I had a hard time with.

Analysis

Each time that I saw a card/question, I would respond and see the result (if I succeeded or failed). This is what I called Review in that part. Some data about my reviews using Anki:


On 2022-10-10, it was a week that I took off because of some personal problems.

In total, I had 4.6k reviews, an average of 361 reviews per week. It is interesting to see that I increased the number of cards while I was getting closer and closer to the exam (in the 2022-12-05 week).


The same thing as the reviews, I increased the time spent over the weeks revisiting the cards. Another interesting metric is the Seconds per Review, that is possible to see that I was able to improve my time to respond to each question (also due to the fact that I was reviewing a lot of questions that I already saw, so it was easier to respond):

Result

Here is my performance and some numbers calculated using the last step of the Analysis:

Some key learnings:

  • I decided to go for quality over quantity, with one certification per semester at maximum. That decision was after I almost was close to NOT passing the AWS Certified Developer exam.
  • Plus, my decision to review more cards for my latest exam allowed me to have my best score yet!
  • Interesting that I'm reducing, after each exam, the time spent per review and improving my score overall.
  • It seems like I found my own process after iterating in the first two exams.

Performance per section for the AWS Certified Machine Learning exam, provided by AWS:
![[Screenshot 2022-12-16 at 13.37.35.png]]

If you intend to study for the exam, I hope that this article could help you to find quality resources and some tips to help you to pass the test!