Apr 17, 2022 Rahul Sharma
These are my notes and preparation strategy for the MLS-C01 exam.
**Reading material**
1. [Udemy course by Chandra Lingam](<https://www.udemy.com/course/aws-machine-learning-a-complete-guide-with-python/>)
2. [Udemy course by Maarek and Kane](<https://www.udemy.com/course/aws-machine-learning/>)
3. [SageMaker FAQ](<https://aws.amazon.com/sagemaker/faqs/>)
4. [SageMaker Deepdive series on YouTube](<https://www.youtube.com/watch?v=uQc8Itd4UTs&list=PLhr1KZpdzukcOr_6j_zmSrvYnLUtgqsZz>)
****
**Practice tests**
1. [Tutorials Dojo](<https://www.udemy.com/course/aws-certified-machine-learning-specialty-practice-exams-amazon/>)
2. [Abhishek Singh](<https://www.udemy.com/course/aws-certified-machine-learning-specialty-full-practice-exams/>)
3. [AWS MLS Practice Test](<https://amazonmr.au1.qualtrics.com/jfe/form/SV_2nt0z0Far6DzAwJ>)
**General information**
1. [AWS ML Ramp-up Guide](<https://d1.awsstatic.com/training-and-certification/ramp-up_guides/Ramp-Up_Guide_Machine_Learning.pdf>)
2. [AWS MLS Exam Guide](<https://d1.awsstatic.com/training-and-certification/docs-ml/AWS-Certified-Machine-Learning-Specialty_Exam-Guide.pdf>)
3. [AWS MLS Exam Readiness](<https://explore.skillbuilder.aws/learn/course/external/view/elearning/27/exam-readiness-aws-certified-machine-learning-specialty>)
I studied for a month, about 2-3 hours per day. I previously did the AWS Solution Architect Associate certification and had a good handle on common services (e.g., EC2, S3, VPC). However, MLS-C01 tests your general ML knowledge more than AWS-specific ones.
I started with Chandra Lingam’s course on Udemy, did a couple practice tests, and found my weak spots. I then turned to the Maarek/Kane course on Udemy to reinforce my learning and finally closed with one practice test per day for a week leading to the day of the exam. Overall, I took 8 practice tests by Tutorials Dojo and Abhishek Singh, scoring 80-90% on all of them.
The exam is heavily focused on SageMaker so make sure you know it like the back of your hand. This includes all the built-in algorithms, model training, hyperparameter tuning, endpoint deployments, and so on. Below is my experience on the type of questions that appeared on the exam.
**More questions on**
1. Glue (Data Catalog, ETL)
2. Confusion matrix (Recall, Precision)
3. AWS ML services (Comprehend, Forecast)
4. Linear Learner
5. Data preprocessing (imputations, unbalanced data)
**Less questions on**
1. Model input parameters (parquet, RecordIO)
2. Hyperparameter tuning
3. Deep Learning (CNN, RNN)
I took the exam from home through Pearson VUE. Logged in 30 minutes before the start time, finished various formalities (e.g., ID proof, system check). Make sure you have a clean working space with no distractions.
I started off slower, flagging several questions for review, and finished the first pass within an hour. I went back to the flagged questions, reviewed them again and was able to finish the complete test with an hour remaining.
As soon as I submitted my exam, I was presented with a small AWS survey post which I got to see the long awaited screen saying PASS in bold letters!