On the path towards technical excellence, one such certification that the Data Scientists at TheDataTeam have taken up recently is the AWS Certified ML Speciality examination.

As pointed out by numerous industry experts, the MLS-C01 exam for AWS Certified ML Speciality is one of the harder cloud certification exams to clear. In order to help aspiring Data Scientists, TheDataTeam has gathered input from its in-house experts to provide a guide to help aspirants prepare for the MLS-C01 examination and accelerate their careers in the field of data science & AI.     

STUDY GUIDE

A note of caution before we begin: the guidance in this post is not meant to be comprehensive, and may not even be adequate, but we believe it would serve as a thorough walkthrough to help streamline your preparations. The reasons why the MLS-C01 exam is considered a hard one to crack is because it has the following characteristics:

  1. Covers diverse subject areas – Topics related to Engineering, Analytics, Modelling and Deployment (including Security & Scaling)
  2. Requires a deep understanding of Data Science concepts
  3. Emphasizes operational machine learning equally as much as the algorithmic contents, including topics like Security, App scaling, Network setup, etc., which a typical MOOCS ML coursework does not really do justice to.
  4. Comprises of questions that test aptitude-based knowledge, as opposed to theoretical understanding alone, i.e. questions were worded subtly, making it tricky
  5. Goes beyond the AWS stack with more than 50% of the questions related to a broader understanding of concepts from the open world of data science, vastly increasing the surface area of knowledge that could make it potentially daunting to an aspirant.

In summary, the exam tests the candidate’s understanding of Data Science from end-to-end, in order to help solve business problems with the most appropriate ML solutions. 

Aspirants should note that the ‘Specialty’ category of the AWS Certification exam is equivalent to the ‘Professional’ category, which is the highest level of certification available right now.

 KEY OBSERVATIONS

Here are a few official details about the exam: MLS-C01 Details (Make sure to read this document carefully before proceeding with the following section)

Some general pointers about the exam description, which are already given in the document:

  1. Every question in the exam sticks to the 3 objectives stated in the document. This gives the exam aspirants a good idea about the template of the questions.
  2. The document recommends 1-2 years of hands-on experience. While experience tends to help fast-track the process of self-preparation, it’s been observed that many others have managed to clear the exam with a few months of preparation, without any prior Data Science experience, though we would strongly advise getting hands-on exposure.
  3. Our in-house experts have pointed out that Multiple Response Questions are inclined to be harder than the rest, but it is also important to note that they award marks across a curve. If an individual was to get 2 out of the 3 possible right answers, they are awarded points corresponding to the 2 correct answers. This can be observed from the official practice exam.
  4. AWS scales the score, like the SAT exam. The exact score for each question is not known, but difficult questions carry more marks (based on the history of correct answers). Read more here.
  5. TheDataTeam’s experts advise exam aspirants to not to leave any question unanswered since no negative marks are awarded.
  6. AWS keeps changing the question bank frequently, Hence, the topic and template of the questions might change.

PREPARATION STEPS

The following steps will help walk aspirants through the preparation process for the exam, based on what worked for our in-house experts:

STEP 1: Testing the waters

Before diving deep into the preparation process, it’s better to understand what you, as a candidate, can expect from this exam. Here is an official sample set of questions. The main exam is at the same level of difficulty.

STEP 2: Start with the Basics

Get familiar with the basics. If fundamental concepts about Data Science are not understood well, then handpick a few courses to help strengthen the areas that need work. Write to us if you need advice on which courses are useful for the same.

As an example, if candidates are new to Data Science or have less practical experience, it is advisable to go through Andrew Ng’s Courses on Coursera. Even for the seasoned practitioners, we recommend the DL Course #2 on Hyperparameter Tuning and Regularization.

Machine Learning by Andrew Ng

Deep Learning Specialization by Andrew Ng (5 Courses)

STEP 3: Official Course Path (Optional)

The official coursework is laid out in detail. As it mostly touches on the topics only relevant to the exam, this is recommended for those who have less practical experience regarding Data Science.

Link to ML Path

[To access the courses, set up an AWS Partner Network (APN) account. Almost all the digital courses (non-classroom) are available for free]

STEP 4: Full Syllabus Courses

There are many full-syllabus courses for this certification exam. The popular ones are on:

  1. aCloudGuru
  2. whizlabs
  3. Udemy

All the courses are adequate and cover most of the syllabus on a high level but not exhaustive. Hence, all three courses are recommended equally, so pick one that works for you.

STEP 5: Amazon Sagemaker Deep Dive

Sagemaker is the main Data Science offering from Amazon. Our experts have pointed out that many questions in the exam, directly or indirectly, were related to Sagemaker. Many online experts advise aspiring candidates to go through the official Sagemaker Product Documentation. However, the PDF is 1300 pages long! The two sources below were found to be sufficient:

  1. a) Concise and Updated Youtube playlist
  2. b) HTML version of the Product Documentation

STEP 6: Practice Exams

This is the most important step of all, from an examination point of view. There are several practice exams available online. A word of caution: aspirants are advised to refrain from using ‘question dumps’ that are available online since these tend to be expensive and usually insufficient or unrepresentative of the actual exam.

STEP 7: Exam Readiness Check

This official course (2 hours) helps candidates by providing a summary of all the topics from the syllabus coupled with mock tests. The assessment from these tests is useful to understand areas that need further deep-dive or renewed re-training.

It is important to not get disheartened by the results of these mock tests. Anecdotal evidence suggests that it is possible that a candidate might get a score between 75 – 85% on the mock exams but still manage to obtain 90%+ on the official MLS-C01 exam.

STEP 8: Official Practice Exams

The course mentioned in Step #7 comes with a practice exam. It has 35 questions, along with a few additional questions under each section. Aspirants should ideally try to attempt these questions multiple times in order to get familiar with the template.

This is the last checkpoint towards overall readiness. If an aspirant finds it difficult to obtain at least 75% in the mock tests, it’s recommended to spend more time in strengthening areas that have emerged as requiring further understanding.

ADDITIONAL EXAM GUIDES

Other MLS-C01 exam experiences are available. Candidates can also search for others online.

Below are a few good examples:

Link#1  Link#2  Link#3  Link#4

TIPS ABOUT THE EXAM THAT OTHER EXPERTS AREN’T TELLING YOU

  1. Templates for most of the questions were similar to the ones from the official and unofficial practice exams. Candidates should invest additional time in analysing each question thoroughly.
  1. There are some algorithms and frameworks that were introduced by AWS through their research teams, like Random Cut Forest, DeepAR, MXNet, etc. Amazon is proud of these offerings, and candidates must expect at least 1 question from these topics.
  2. Some unexpected questions that were present in the exams were from Security (NACL, VPC Endpoints, etc), operationalization including deployment, and about new services that were introduced recently (such as Persona, Forecast (DeepAR), Macie, etc).
  3. The Specialty and Professional exam costs $300 + 18% Service Tax (as on date). However, there is a way to reduce this cost by half! All a candidate has to do is take the Level 1 AWS Cloud Certified Professional (CCP) exam first, as it only costs $100 + 18% service tax. Once the candidate clears this rather easy exam, they’ll get a 50% discount voucher and a FREE practice Exam voucher, which would have otherwise cost an additional $50.
  4. Since Indians are non-native English speakers, aspiring Data Scientists from the subcontinent can avail an additional 30 minutes for every AWS exam. This requires the candidate to enable a certain option on the VUE portal before booking the exam. Note: Since everyone attempting the exam gets 165 mins for the MLS-C01 and 90 mins for the CCP, it might very well be that the additional 30 mins are not required in most cases.
  5. Answers for one question might be linked to another question that may appear later in the exam. Candidates can use the bookmark feature to revisit questions quickly instead of wasting time searching for the question again.
  6. There is an adequate amount of time to complete the exam, with experts stating that individuals appearing for the exam have enough time to go through each question twice.
  7. Make sure to not miss any keyword in the question. These keywords are tied to the provided answer options. For example, if the question says, “optimize cost”, then discard the options that might solve the problem with a higher cost. Please note that multiple options might solve the given case, but only one correct answer might exist based on the given keywords.
  8. The exam window closes 15 mins before the scheduled time. This is because it takes at least 15 mins to complete the pre-requisites, which includes capturing images of surroundings, photo ID, etc.
  9. The result is displayed on the screen immediately after completing the procedure, but the scores and certificates take 1-2 days.
  10. Pro Tip: Don’t book the exam for a Sunday early morning. Our in-house experts pointed out that one occasion, their proctor had slept off and they had to request for a reschedule! The proctor then woke up and started the exam after 45 mins.

For everyone who is about to attempt the MLS-C01 examination, we wish them all the best and hope this guide comes in handy during your preparations.

Good luck and Godspeed.

TheDataTeam

Leave a Reply

Close Bitnami banner
Bitnami