Clearing AWS ML Specialty Certification in just 6 days of study with a cheat sheet

Harsh Vardhan Solanki
5 min readJan 28, 2021

--

AWS Certified ML Specialty Certificate| Source: AWS

Day 1: Exploring what you are getting into

Day 2: Starting the Udemy Course for AWS ML Specialty

Day 3: Finishing the Udemy Course for AWS ML Specialty

Day 4: Start giving practice exams

Day 5: Prepare notes of the important topics to go through before exam day

Day 6: Recap important topics and just trust your gut

Day 1: Exploring what you are getting into

It is often a good practice (specially for a data scientist/machine learning engineer) to explore the data first, get a mental picture of what you are trying to build and what it will achieve. Similarly, you should know what to expect from the exam. The exam is basically divided into 2 parts:

Generic ML + ML on AWS

  • You would ideally be tested on 65 questions in total
  • Questions are all situational based and revolve around all aspects of what ML engineers is expected to do with a given business problem such as :
  1. Exploratory Data Analysis
  2. Feature Engineering
  3. Data Preparation
  4. Model Tuning
  5. Inferences
  6. Infrastructure on AWS

Services that should be a must know for this certification:

  • Sagemaker
  • Kinesis
  • EMR
  • Lambda
  • IoT Greengrass
  • Glue

If you are just getting started and are a refresher, you might need help in preparation of the first part (Generic ML) too. Following resources would be more than enough to help you prepare for that:

Training Material Links :

Introduction to Machine Learning — https://www.coursera.org/learn/machine-learning/home/welcome

Introduction to Deep Learning -

https://www.coursera.org/learn/neural-networks-deep-learning/home/welcome

The above resources can be skipped if you already know about Machine Learning or have worked as a ML engineer.

Day 2: Starting the Udemy Course for AWS ML Specialty

Note: I am in no way associated with Udemy/Sundog/Stephane/Frank. It’s just my opinion and experience on the basis of which, I am suggesting their course. (It’s cheap too compared to other counterparts)

For some reason, you do not want to take the above course, here is an another alternative which is more or less same as this one (even better, some might say):

Day 3: Finishing the Udemy Course for AWS ML Specialty

Well, the above courses are 9 and 13 hours respectively. So, if you cannot sit in a stretch without consuming lot of caffeine like me, you will need day 2 to finish the course.

I would recommend taking two days (atleast) to imbibe what you will be learning from the course.

Day 4: Start giving practice exams

The courses did not help me as much as practice exams did. This is because I had enough hands on experience due to the nature of my work where I have either used those services or I had read those concepts earlier. So, I skipped past few sections of the course (not recommended if you are just getting started with ML)

Now that you have learnt enough from the course, it is time to put it to test. And, following practice exams are more or less similar to what you will get to see in the certification itself. A huge shoutout to these amazing instructors/gurus for preparing such a wonderful set of questions to test our knowledge:

  • Both the courses come with few example practice questions at the end. The quality and quantity of these questions is way more than that of Udemy course.
  • You can purchase following practice exams separately from the course and these are at par with what you will get in cloudguru. I happen to be a person who likes to be more prepared than more sorry. So, I ended up taking all three practice exams from Udemy and Cloudguru (it did wonders for me)

Day 5: Prepare notes of the important topics to go through before exam day

By now, you might have already prepared your own notes from the practice exam and the course. However, I would recommend making a separate short version of those notes (think: something like flash cards) for the areas you are lacking in (practice exams will help you know this)

For me, it was confusion matrix (this always confused me) but I figured a way to understand this once and for all with a trick.

Apart from that, I struggled with security related questions too. So, I ended up writing all the major concepts in such flash cards to go through before the exam.

Day 6: Recap important topics and just trust your gut

If you have done all of the above, you are already prepared for the examination. I was nervous too but apparently I was prepared even though I did not trusted myself.

So, do not make the same mistake. However, if you are not very sit-idle-at-a-place kind, you might need some mental preparation for the idea of around 3 and half hour long examinations.

Some Additional Pointers:

  • Be prepared to sit for 3.5 hours long without even moving (you can stretch if you are under the web cam view though)
  • Eat something but not too heavy before examination
  • Schedule it when you will most energetic
  • Use Flag option during examination and think hard enough on these questions at last
  • Take your time (you will have plenty). I was making this mistake earlier of finishing the exams in around half the time allotted (this led to lot of silly mistakes). I thought I will not be able to finish it on time, so I was rushing through the questions missing few key terms in the question.
  • Use elimination and try to answer before seeing the options

I will be writing another piece containing the cheat sheet of the topics that you are likely to see in this examination.

With this, I bid adieu to you. All the very best! Go get yours! It feels amazing to have these kind of certifications.

P.S: Please do reach out to me for my own personal notes along with practice questions that I am very proud to say that will be enough alone to clear the certification.

--

--

Harsh Vardhan Solanki
Harsh Vardhan Solanki

Written by Harsh Vardhan Solanki

Machine Learning Engineer | Here to share my thoughts on all things data science.

Responses (1)