{"id":1487,"date":"2020-07-10T12:02:47","date_gmt":"2020-07-10T10:02:47","guid":{"rendered":"https:\/\/blog.besharp.it\/?p=1487"},"modified":"2024-02-02T11:56:31","modified_gmt":"2024-02-02T10:56:31","slug":"aws-machine-learning-specialty-how-i-got-certified-in-ten-days-as-a-devops-engineer","status":"publish","type":"post","link":"https:\/\/blog.besharp.it\/aws-machine-learning-specialty-how-i-got-certified-in-ten-days-as-a-devops-engineer\/","title":{"rendered":"AWS Machine Learning Specialty: How I got certified in ten days as a Devops Engineer"},"content":{"rendered":"<h2 class=\"p1\"><span class=\"s1\">Introduction<\/span><\/h2>\n<p class=\"p2\"><span class=\"s1\">He did it!<\/span><\/p>\n<p class=\"p2\"><span class=\"s1\"><strong>Our Cloud Expert Alessandro Gaggia got his sixth (!) AWS Certification<\/strong> (the 58th AWS Certification for beSharp): the <strong>AWS Certified Machine Learning Specialty!\u00a0<\/strong><\/span><\/p>\n<p class=\"p2\"><span class=\"s1\">Alessandro is considered a backbone of our company: he joined the team as a Front-end developer back in 2012, a few months after beSharp&#8217;s establishment. Today, he is the dean of our development team. He first approached Machine Learning one year ago while participating in the AWS DeepRacer League held during the AWS RE: Mars in Las Vegas. From that moment onwards, it became a real passion for him. After the rush, which took him towards the certification, Alessandro threw himself into a precious retrospective study on the winning path that led him to obtain this excellent result in a few days. So it&#8217;s time to leave the floor. Ready to become AWS Certified?<\/span><\/p>\n<blockquote class=\"twitter-tweet\" data-width=\"500\" data-dnt=\"true\">\n<p lang=\"en\" dir=\"ltr\">Many congrats to our <a href=\"https:\/\/twitter.com\/hashtag\/CloudExpert?src=hash&amp;ref_src=twsrc%5Etfw\">#CloudExpert<\/a> <a href=\"https:\/\/twitter.com\/Balubor?ref_src=twsrc%5Etfw\">@Balubor<\/a>! \ud83c\udf89<br \/>So <a href=\"https:\/\/twitter.com\/hashtag\/Proud2beCloud?src=hash&amp;ref_src=twsrc%5Etfw\">#Proud2beCloud<\/a>! <a href=\"https:\/\/twitter.com\/hashtag\/AWSCertified?src=hash&amp;ref_src=twsrc%5Etfw\">#AWSCertified<\/a> <a href=\"https:\/\/twitter.com\/hashtag\/MachineLearning?src=hash&amp;ref_src=twsrc%5Etfw\">#MachineLearning<\/a> <a href=\"https:\/\/twitter.com\/hashtag\/AWS?src=hash&amp;ref_src=twsrc%5Etfw\">#AWS<\/a> <a href=\"https:\/\/twitter.com\/awscloud?ref_src=twsrc%5Etfw\">@awscloud<\/a> <a href=\"https:\/\/twitter.com\/AWSCertifiedBot?ref_src=twsrc%5Etfw\">@AWSCertifiedBot<\/a> <a href=\"https:\/\/t.co\/viMskr5nL1\">https:\/\/t.co\/viMskr5nL1<\/a><\/p>\n<p>&mdash; beSharp (@beSharp_spa) <a href=\"https:\/\/twitter.com\/beSharp_spa\/status\/1269931019383250945?ref_src=twsrc%5Etfw\">June 8, 2020<\/a><\/p><\/blockquote>\n<p><script async src=\"https:\/\/platform.twitter.com\/widgets.js\" charset=\"utf-8\"><\/script><\/p>\n<h2 class=\"p1\"><span class=\"s1\">My personal experience<\/span><\/h2>\n<p class=\"p4\"><span class=\"s2\"><strong>AWS Certified Machine Learning\u00a0 &#8211; Specialty<\/strong> is an advanced certification a bit different from the others, because it is the only one which focuses on specific sector knowledge not strictly tied to AWS services. In fact, in order to pass the exam and obtain the certification, it\u2019s fundamental being able to <strong>recognize, analyze and optimize different machine learning problems<\/strong> starting from use cases\u2019 descriptions, without them being exclusively linked to peculiar AWS\u2019 solutions.<\/span><\/p>\n<p class=\"p4\"><span class=\"s1\">This is my point of view and it diverges a little from what I have read in many articles, where authors consider SageMaker as the main focus of the exam. I believe, instead, that this exam tested mainly me being familiar with Machine Learning in general.<\/span><\/p>\n<p class=\"p4\"><span class=\"s1\">On the other hand, many testimonials &#8211; given by expert data scientists &#8211; say that, before taking this certification, many years of on-the-field training are required. Personally, I don\u2019t believe this is entirely true: experience always plays a fundamental role however, like many other certifications I\u2019ve obtained, what you really need is dedication to study, a strong desire to deep dive in all the arguments and logic ability to make it through. Also some understanding of Big Data concepts can help.\u00a0\u00a0<\/span><\/p>\n<p class=\"p4\"><span class=\"s2\">Disclaimer: information gathered here rep\u00f2resent my personal experience on preparing the exam and it is by no means to be intended as a substitute, more over exhaustive, of studying material proposed by Amazon Web Services. I believe though, that they are an excellent reference material to follow in order to optimize your learning path, containing references to many different sites about Machine Learning and on AWS\u2019 FAQ.\u00a0\u00a0<\/span><\/p>\n<p class=\"p4\"><span class=\"s2\">Be also extra careful about the different <strong>questionnaires available for testing<\/strong>, for which you\u2019ll get the link, in particular the Certbolt executable, which allows to simulate a complete test of 65 questions with solutions.<\/span><\/p>\n<p class=\"p4\"><span class=\"s2\">With a study full immersion session of about 8 hours a day, certification can be achieved in 1 days\u2026 or at least, this was my strategy, which gave me the ability to pass the exam with a score of 800\/1000 on a minimum of 750\/1000.\u00a0<\/span><\/p>\n<h2 class=\"p5\"><span class=\"s2\">Exam\u2019s topics<\/span><\/h2>\n<p class=\"p4\"><span class=\"s2\">The exam is <strong>170 minutes <\/strong>long for<strong> 65 questions<\/strong>, personally, I managed to finish in 9 minutes, which gave me plenty of extra time for checking. One positive note is that, unlike other advanced certifications, this one has <strong>short use cases, which means less time to read and more time to think.\u00a0<\/strong><\/span><\/p>\n<p class=\"p4\"><span class=\"s2\">I\u2019ve put a list of all the main topics i found in both exam readiness test and in the official one:<\/span><\/p>\n<ul class=\"ul1\">\n<li class=\"li4\"><span class=\"s2\">Given a ML problem, use a combination of AWS managed services to solve it in the most fast and efficient way possible. In this case we have the following topics:\u00a0<\/span><span class=\"s1\"><br \/>\n<\/span><\/p>\n<ul class=\"ul2\">\n<li class=\"li4\"><b><\/b><span class=\"s2\"><b>Kinesis Data Stream, Analytics, Firehose.<\/b><\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><b><\/b><span class=\"s2\"><b>S3, KMS, DynamoDB <\/b>(just remember what they are and what they do)<\/span><span class=\"s1\"><b><br \/>\n<\/b><\/span><\/li>\n<li class=\"li4\"><b><\/b><span class=\"s2\"><b>AWS ElasticSearch con Kibana, Splunk, AWS Quicksight <\/b>(for Business intelligence &#8211; a.k.a <b>BI<\/b>)<\/span><span class=\"s1\"><b><br \/>\n<\/b><\/span><\/li>\n<li class=\"li4\"><b><\/b><span class=\"s2\"><b>AWS Athena, AWS Redshift<\/b><\/span><span class=\"s1\"><b><br \/>\n<\/b><\/span><\/li>\n<li class=\"li4\"><b><\/b><span class=\"s2\"><b>AWS EMR with Spark<\/b><\/span><span class=\"s1\"><b><br \/>\n<\/b><\/span><\/li>\n<li class=\"li4\"><b><\/b><span class=\"s2\"><b>AWS Glue<\/b><\/span><span class=\"s1\"><b><br \/>\n<\/b><\/span><\/li>\n<li class=\"li4\"><b><\/b><span class=\"s2\"><b>AWS Translate, Transcribe, Lex, Polly, DeepLens, Rekognition, Comprehend, etc.\u00a0<\/b><\/span><span class=\"s1\"><b><br \/>\n<\/b><\/span><\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<p class=\"p4\"><span class=\"s2\">In some cases we also have a combination of Managed Services and SageMaker, but in these specific cases it\u2019s usually trivial to exclude wrong answers.<\/span><\/p>\n<ul class=\"ul1\">\n<li class=\"li4\"><span class=\"s2\">Given a ML problem, being able to manage the following scenarios:<\/span><span class=\"s1\"><br \/>\n<\/span><\/p>\n<ul class=\"ul2\">\n<li class=\"li4\"><span class=\"s2\">Data cleaning with appropriate techniques like: <b>Filtering, Transforming, Scaling, Extracting<\/b><\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Manage <b>Overfitting\/Underfitting<\/b>, understanding when we have to work with: <b>Learning Rate, Batch Size, Oversampling, Regularization, Dropout Rate, Feature Increasing, Denoising, Normalization, Epoch Time, etc.<\/b><\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Evaluate the appropriate score for a specific model, keeping in mind its typology and Business requests; learn how to choose between: <b>RMSE, R2, F1, Recall, Accuracy, Precision, AUS-ROC, etc.<\/b><\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Manage <b>Hyperparameters<\/b> tuning both on Sagemaker and in general; you need to see: <b>Learning Rate, Epoch Time, L1 e L2 Regularization, Adam, Random and Stocastic Optimization, alpha, beta <\/b>and<b> gamma parameters <\/b>in different contexts.<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><strong><span class=\"s2\">Manage deploy on Sagemaker.<\/span><\/strong><span class=\"s1\"><br \/>\n<\/span><\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<p class=\"p4\"><span class=\"s2\">To resume all the projects in details, some \u201cmust\u201d are:<\/span><\/p>\n<ul class=\"ul1\">\n<li class=\"li4\"><span class=\"s2\">Creation of a pipeline using Kinesis Stream, Firehose and Analytics in combination with Amazon Athena or Elastic Search<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<\/ul>\n<ul class=\"ul1\">\n<li class=\"li4\"><span class=\"s2\">Solve a Overfitting problem<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Solve a Tuning problem<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Solve a Scoring problem<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Solve a Confusion matrix problem<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Use a combination of managed AWS ML services to solve a problem in a easy and efficient way<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><b><\/b><span class=\"s2\"><b>Manage different deploy scenarios on Sagemaker<\/b><\/span><span class=\"s1\"><b><br \/>\n<\/b><\/span><\/li>\n<\/ul>\n<p class=\"p4\"><span class=\"s2\">This list covers the majority of questions you\u2019ll likely to see in an exam. Following I will propose a study guide, making it possible to manage all the topics in a tidy and organic way.<\/span><\/p>\n<h2 class=\"p1\"><span class=\"s1\">Topics to study: a lineup<\/span><\/h2>\n<p class=\"p2\"><span class=\"s1\">Reference links about topics here described are mostly from AWS official documentation, in particular for SageMaker and others managed services. Also, regarding topics more related to Machine Learning, I want to propose some sites that I found particularly complete, well explained and exhaustive.\u00a0<\/span><\/p>\n<ul class=\"ul1\">\n<li class=\"li6\"><span class=\"s1\">What is machine learning: how is different from Artificial Intelligence and Deep Learning: <span class=\"s5\"><a href=\"https:\/\/machinelearningmastery.com\/machine-learning-for-programmers\/\">https:\/\/machinelearningmastery.com\/machine-learning-for-programmers\/<\/a><br \/>\n<\/span><\/span><a href=\"https:\/\/www.geeksforgeeks.org\/difference-between-artificial-intelligence-vs-machine-learning-vs-deep-learning\/?ref=rp\">https:\/\/www.geeksforgeeks.org\/difference-between-artificial-intelligence-vs-machine-learning-vs-deep-learning\/?ref=rp<\/a><\/li>\n<\/ul>\n<ul class=\"ul1\">\n<li class=\"li8\"><span class=\"s8\">ML pipeline in general and on AWS: <a href=\"https:\/\/medium.com\/slalom-data-analytics\/mlops-part-2-machine-learning-pipeline-automation-with-aws-1ca10348239e\"><span class=\"s9\">https:\/\/medium.com\/slalom-data-analytics\/mlops-part-2-machine-learning-pipeline-automation-with-aws-1ca10348239e<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">How we proceed on managing a ML problem:<br \/>\n<\/span><a href=\"https:\/\/towardsdatascience.com\/how-to-approach-a-machine-learning-problem-3fe843fd1166\">https:\/\/towardsdatascience.com\/how-to-approach-a-machine-learning-problem-3fe843fd1166<\/a><a href=\"https:\/\/towardsdatascience.com\/task-cheatsheet-for-almost-every-machine-learning-project-d0946861c6d0\">https:\/\/towardsdatascience.com\/task-cheatsheet-for-almost-every-machine-learning-project-d0946861c6d0<\/a><\/li>\n<\/ul>\n<ul class=\"ul1\">\n<li class=\"li6\"><span class=\"s1\">Machine Learning on AWS introduction: Managed Services (focus on offered services)<br \/>\n<\/span><\/p>\n<ul class=\"ul2\">\n<li class=\"li8\"><span class=\"s8\">AWS GLUE: <a href=\"https:\/\/aws.amazon.com\/it\/glue\/faqs\/\"><span class=\"s9\">https:\/\/aws.amazon.com\/it\/glue\/faqs\/<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">AWS DATAPIPELINE: <a href=\"https:\/\/aws.amazon.com\/it\/datapipeline\/faqs\/\"><span class=\"s9\">https:\/\/aws.amazon.com\/it\/datapipeline\/faqs\/<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">AWS AUGMENTED AI: <a href=\"https:\/\/aws.amazon.com\/it\/augmented-ai\/faqs\/\"><span class=\"s9\">https:\/\/aws.amazon.com\/it\/augmented-ai\/faqs\/<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">AWS DEEP LENS: <a href=\"https:\/\/aws.amazon.com\/it\/deeplens\/faqs\/\"><span class=\"s9\">https:\/\/aws.amazon.com\/it\/deeplens\/faqs\/<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">AWS TRANSLATE: <a href=\"https:\/\/aws.amazon.com\/it\/translate\/faqs\/\"><span class=\"s9\">https:\/\/aws.amazon.com\/it\/translate\/faqs\/<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">AMAZON TRANSCRIBE: <a href=\"https:\/\/aws.amazon.com\/it\/transcribe\/faqs\/\"><span class=\"s9\">https:\/\/aws.amazon.com\/it\/transcribe\/faqs\/<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">AMAZON TEXTRACT: <a href=\"https:\/\/aws.amazon.com\/it\/textract\/faqs\/\"><span class=\"s9\">https:\/\/aws.amazon.com\/it\/textract\/faqs\/<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">AWS REKOGNITION: <a href=\"https:\/\/aws.amazon.com\/it\/rekognition\/faqs\/\"><span class=\"s9\">https:\/\/aws.amazon.com\/it\/rekognition\/faqs\/<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">AWS POLLY: <a href=\"https:\/\/aws.amazon.com\/it\/polly\/faqs\/\"><span class=\"s9\">https:\/\/aws.amazon.com\/it\/polly\/faqs\/<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">AWS PERSONALIZE: <a href=\"https:\/\/aws.amazon.com\/it\/personalize\/faqs\/\"><span class=\"s9\">https:\/\/aws.amazon.com\/it\/personalize\/faqs\/<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">AMAZON LEX: <a href=\"https:\/\/aws.amazon.com\/it\/lex\/faqs\/\"><span class=\"s9\">https:\/\/aws.amazon.com\/it\/lex\/faqs\/<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">AMAZON KENDRA: <a href=\"https:\/\/aws.amazon.com\/it\/kendra\/faqs\/\"><span class=\"s9\">https:\/\/aws.amazon.com\/it\/kendra\/faqs\/<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">AMAZON FORECAST: <a href=\"https:\/\/aws.amazon.com\/it\/forecast\/faqs\/\"><span class=\"s9\">https:\/\/aws.amazon.com\/it\/forecast\/faqs\/<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">AMAZON COMPREHEND: <a href=\"https:\/\/aws.amazon.com\/it\/comprehend\/faqs\/\"><span class=\"s9\">https:\/\/aws.amazon.com\/it\/comprehend\/faqs\/<\/span><\/a><br \/>\n<\/span><\/li>\n<\/ul>\n<\/li>\n<li class=\"li8\"><span class=\"s8\">ML Supervised: <a href=\"https:\/\/towardsdatascience.com\/types-of-machine-learning-algorithms-you-should-know-953a08248861\"><span class=\"s9\">https:\/\/towardsdatascience.com\/types-of-machine-learning-algorithms-you-should-know-953a08248861<\/span><\/a><br \/>\n<\/span><\/p>\n<ul class=\"ul2\">\n<li class=\"li6\"><span class=\"s1\">Linear Regression<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Forecasting<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Classification<br \/>\n<\/span><\/li>\n<\/ul>\n<\/li>\n<li class=\"li8\"><span class=\"s8\">ML Unsupervised: <a href=\"https:\/\/www.guru99.com\/unsupervised-machine-learning.html\"><span class=\"s9\">https:\/\/www.guru99.com\/unsupervised-machine-learning.html<\/span><\/a><br \/>\n<\/span><\/p>\n<ul class=\"ul2\">\n<li class=\"li6\"><span class=\"s1\">Clustering<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Anomaly Detection<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Topic modeling<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Machine translation<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Reinforcement Learning<br \/>\n<\/span><\/li>\n<\/ul>\n<\/li>\n<li class=\"li8\"><span class=\"s8\">Sagemaker Ground Truth: <span class=\"s9\">https:\/\/aws.amazon.com\/it\/sagemaker\/groundtruth\/faqs\/<\/span><br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">Datalake concepts and S3: <a href=\"https:\/\/aws.amazon.com\/it\/big-data\/datalakes-and-analytics\/what-is-a-data-lake\/?nc=sn&amp;loc=2\"><span class=\"s9\">https:\/\/aws.amazon.com\/it\/big-data\/datalakes-and-analytics\/what-is-a-data-lake\/?nc=sn&amp;loc=2<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Data cleaning and formatting:<br \/>\n<\/span><\/p>\n<ul class=\"ul2\">\n<li class=\"li6\"><span class=\"s1\">Nominal and categorical data<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Unclean data<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Outliers<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Missing data<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Graphics with Pandas<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Correlation Matrix<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Overfitting and Underfitting<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Data splitting<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Simple hold out<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">K Fold<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">K Fold random shuffle<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Stratified K Fold<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Leave one out<br \/>\n<\/span><\/li>\n<\/ul>\n<\/li>\n<li class=\"li6\"><span class=\"s1\">Tools for Supervised\/Unsupervised analysis:<br \/>\n<\/span><\/p>\n<ul class=\"ul2\">\n<li class=\"li8\"><span class=\"s8\">ML Supervised: <a href=\"https:\/\/towardsdatascience.com\/metrics-to-evaluate-your-machine-learning-algorithm-f10ba6e38234\"><span class=\"s9\">https:\/\/towardsdatascience.com\/metrics-to-evaluate-your-machine-learning-algorithm-f10ba6e38234<\/span><\/a><br \/>\n<\/span><\/p>\n<ul class=\"ul3\">\n<li class=\"li6\"><span class=\"s1\">RMSE, MSE, R2<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Accuracy, F1 Score, Precision, Recall, TNR, AUC-ROC<br \/>\n<\/span><\/li>\n<\/ul>\n<\/li>\n<li class=\"li8\"><span class=\"s8\">ML Unsupervised: <a href=\"https:\/\/www.guavus.com\/technical-blog\/unsupervised-machine-learning-validation-techniques\/\"><span class=\"s9\">https:\/\/www.guavus.com\/technical-blog\/unsupervised-machine-learning-validation-techniques\/<\/span><\/a><br \/>\n<\/span><\/p>\n<ul class=\"ul3\">\n<li class=\"li6\"><span class=\"s1\">Internal validation<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">External validation<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Twin-Sample Validation<br \/>\n<\/span><\/li>\n<\/ul>\n<\/li>\n<li class=\"li6\"><span class=\"s1\">ML Deep Learning: know at least some bits<br \/>\n<\/span><\/li>\n<\/ul>\n<\/li>\n<li class=\"li6\"><span class=\"s1\">Sagemaker: Training Job, how to manage and create configurations: <a href=\"https:\/\/docs.aws.amazon.com\/sagemaker\/latest\/dg\/train-model.html\"><span class=\"s5\">https:\/\/docs.aws.amazon.com\/sagemaker\/latest\/dg\/train-model.html<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Sagemaker: possible algorithm solutions<br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">Sagemaker: algorithms in details: <a href=\"https:\/\/docs.aws.amazon.com\/sagemaker\/latest\/dg\/algos.html\"><span class=\"s9\">https:\/\/docs.aws.amazon.com\/sagemaker\/latest\/dg\/algos.html<\/span><\/a><br \/>\n<\/span><\/p>\n<ul class=\"ul2\">\n<li class=\"li6\"><span class=\"s1\">Blazing Text<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Deep AR<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Factorization Machine<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Image Classification<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">IP Insight<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">K-Means<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">KNN<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">LDA<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Linear Learner<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Neural Topic Model<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Object2Vect<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Object Detection<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Principal Component analysis<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Random Cut Forest<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Semantic Segmentation<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Seq2Seq<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">XGBoost<br \/>\n<\/span><\/li>\n<\/ul>\n<\/li>\n<li class=\"li8\"><span class=\"s8\">Sagemaker: tuning: <a href=\"https:\/\/towardsdatascience.com\/demystifying-model-training-tuning-f4e6b46e7307\"><span class=\"s9\">https:\/\/towardsdatascience.com\/demystifying-model-training-tuning-f4e6b46e7307<\/span><\/a><br \/>\n<\/span><\/p>\n<ul class=\"ul2\">\n<li class=\"li6\"><span class=\"s1\">Feature extraction<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Numeric Transformation<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Binning<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Scaling<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Categorical or nominal data<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">How to manage parameters<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">How to manage hyperparameter<br \/>\n<\/span><\/li>\n<\/ul>\n<\/li>\n<li class=\"li8\"><span class=\"s8\">Sagemaker: deploy: <a href=\"https:\/\/docs.aws.amazon.com\/sagemaker\/latest\/dg\/deploy-model.html\"><span class=\"s9\">https:\/\/docs.aws.amazon.com\/sagemaker\/latest\/dg\/deploy-model.html<\/span><\/a><br \/>\n<\/span><\/p>\n<ul class=\"ul2\">\n<li class=\"li6\"><span class=\"s1\">Deploy by yourself<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">Deploy with Sagemaker<br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">How to deploy<br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">Blue\/Green Deploy: <span class=\"s9\">https:\/\/docs.aws.amazon.com\/whitepapers\/latest\/wellarchitected-machine-learning-lens\/bluegreen-deployments.html<\/span><br \/>\n<\/span><\/li>\n<li class=\"li6\"><span class=\"s1\">A\/B Deploy:<br \/>\n<\/span><\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<p class=\"p9\"><span class=\"s6\">https:\/\/docs.aws.amazon.com\/whitepapers\/latest\/wellarchitected-machine-learning-lens\/ab-testing.html<\/span><\/p>\n<ul class=\"ul1\">\n<li style=\"list-style-type: none;\">\n<ul class=\"ul2\">\n<li class=\"li8\"><span class=\"s8\">Canary Deploy: <span class=\"s9\">https:\/\/docs.aws.amazon.com\/whitepapers\/latest\/wellarchitected-machine-learning-lens\/canary-deployment.html<\/span><br \/>\n<\/span><\/li>\n<\/ul>\n<\/li>\n<li class=\"li8\"><span class=\"s8\">Batch Inference: <a href=\"https:\/\/docs.aws.amazon.com\/sagemaker\/latest\/dg\/batch-transform.html\"><span class=\"s9\">https:\/\/docs.aws.amazon.com\/sagemaker\/latest\/dg\/batch-transform.html<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">Online Inference: <a href=\"https:\/\/docs.aws.amazon.com\/sagemaker\/latest\/dg\/inference-pipeline-real-time.html\"><span class=\"s9\">https:\/\/docs.aws.amazon.com\/sagemaker\/latest\/dg\/inference-pipeline-real-time.html<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">Online Vs Batch: <a href=\"https:\/\/mlinproduction.com\/batch-inference-vs-online-inference\/\"><span class=\"s9\">https:\/\/mlinproduction.com\/batch-inference-vs-online-inference\/<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">Sagemaker: logging and Concept Drift: <a href=\"https:\/\/docs.aws.amazon.com\/sagemaker\/latest\/dg\/monitoring-overview.html\"><span class=\"s9\">https:\/\/docs.aws.amazon.com\/sagemaker\/latest\/dg\/monitoring-overview.html<\/span><\/a><br \/>\n<\/span><\/li>\n<li class=\"li8\"><span class=\"s8\">Kinesis Data Stream VS Firehose: <a href=\"https:\/\/aws.amazon.com\/it\/kinesis\/data-streams\/faqs\/?nc=sn&amp;loc=5\"><span class=\"s9\">https:\/\/aws.amazon.com\/it\/kinesis\/data-streams\/faqs\/?nc=sn&amp;loc=5<\/span><\/a><br \/>\n<\/span><\/li>\n<\/ul>\n<p class=\"p7\"><span class=\"s6\"><a href=\"https:\/\/aws.amazon.com\/it\/kinesis\/data-firehose\/faqs\/\">https:\/\/aws.amazon.com\/it\/kinesis\/data-firehose\/faqs\/<\/a><\/span><\/p>\n<h2 class=\"p5\"><span class=\"s2\">Cheats, tips &amp; tricks<\/span><\/h2>\n<ol class=\"ol1\">\n<li class=\"li4\"><span class=\"s2\">Typically, in the case of <b>Simple Hold Out,<\/b> standard values are <b>80\/10\/10<\/b> or <b>70\/15\/15<\/b>.<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">If in some questions you find something related to <b>historical data<\/b> you\u2019ll have to work with <b>supervised algorithms<\/b>.<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">What are the benefits of <b>Random<\/b> and <b>Bayesian<\/b> <b>Optimizer<\/b> for hyperparameters in Sagemaker compared to <b>Grid Optimizer or manual? They better explore parameters space better verifying unusual combinations.<\/b><\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Random optimizer is faster than Bayesian, but the second one is more precise.<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Confusion matrix can also be <b>NxN in dimension<\/b>.<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Term Frequency &#8211; Inverse Document Frequency: a high value means a rare term.<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Oversampling and undersampling <b>are not useful for regression (both logistic and linear)<\/b><\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">When we have a strong unbalanced class, <b>class probability threshold can be put from 0.5 to a higher value for the unbalanced class.<\/b><\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Oversampling <b>must be done after splitting to avoid data bleeding.<\/b><\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">To <b>exit from local minima<\/b> you need to reduce<b> batch size<\/b>\u00a0and <b>also learning rate to reduce oscillatory effect of a small batch size promoting a better convergence.<\/b><\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">For oversampling a basic approach in questions is: <b>GAN is better than SMOTE which is better than manual.<\/b><\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">When we have missing values <b>a viable solution is applying machine learning techniques to replace values<\/b>, in general, is a good approach.<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">When I have questions regarding <b>recommendation engine or in general how to predict users\u2019 choices<\/b>\u00a0if we have <b>collaborative filtering <\/b>as a possible answer than it\u2019s usually a good answer.<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">If you must decide: <b>AWS Glue is more batch oriented, Kinesis Firehose is for streaming real time data.<\/b><\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">If you have forecasting climate problems, you generally have to use <b>Linear Regression (if there aren\u2019t more fitting algorithms, this is usually sufficient)<\/b>.<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Firehose input lambda has a timeout of <b>3 sec, so if you want to execute longer and complex input operations, this value is not sufficient and must be increased.<\/b><\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Scatter Plot is best used for 2 dimension analysis, Histograms for 1 dimension.<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Logistic Regression: <b>is used for binary classification, don\u2019t be fooled by its name!<\/b><\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">NTM: neural topic model: is used to group topics into categories.<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Semantic Segmentation: helps to find high level features by making a per pixel analysis of an image.<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li9\"><span class=\"s1\"><span style=\"font-weight: 400;\">Firehose is not particularly useful if data is already in S3, it takes data from a stream!<\/span><\/span><\/li>\n<li class=\"li9\"><span class=\"s7\"><span style=\"font-weight: 400;\">For deep learning and overfitting problems, use these schemas as a valid reference: <\/span> <\/span><span class=\"s1\"><br \/>\n<\/span><\/p>\n<p style=\"padding-left: 40px;\"><b><\/b><span class=\"s7\"><b>&lt; DROPOUT + &gt; REGULARIZATION + &gt; FEATURE INCREASE = &lt; OVERFITTING<\/b><\/span><span class=\"s1\"><br \/>\n<\/span><\/p>\n<p class=\"p10\" style=\"padding-left: 120px;\"><span class=\"s7\">Sono i parametri da modificare per ridurre l\u2019overfitting. Inoltre:<\/span><\/p>\n<p style=\"padding-left: 40px;\"><b><\/b><span class=\"s7\"><b>&gt; DROPOUT = &gt; NOISINESS<\/b><\/span><span class=\"s1\"><br \/>\n<\/span><\/p>\n<\/li>\n<li class=\"li9\"><span class=\"s1\"><span style=\"font-weight: 400;\">When you\u2019re using Kinesis Data Stream there is a special variant <\/span><b>video stream<\/b><span style=\"font-weight: 400;\"> that is used <\/span><b>in combination with Rekognition Video in a simple and clear way<\/b><span style=\"font-weight: 400;\">. There is a specific use case in AWS guides.<\/span><\/span><\/li>\n<li class=\"li9\"><span class=\"s1\"><span style=\"font-weight: 400;\">AWS Comprehend <\/span><b>is not only<\/b><span style=\"font-weight: 400;\"> for <\/span><b>Sentiment Analysis<\/b><span style=\"font-weight: 400;\"> but also for <\/span><b>Topic Analysis<\/b><span style=\"font-weight: 400;\">, <\/span><b>Key Extraction,<\/b><span style=\"font-weight: 400;\"> and <\/span><b>Entity Recognition<\/b><span style=\"font-weight: 400;\">, also it uses NLP, so in case you\u2019ll have the same topics in a question, you can better understand if you need to use NLP.<\/span><\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">L1 and L2 are regularization terms used to reduce overfitting making the algorithms more conservative when you have a dataset with many features.<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Cross Validation is based on hyperparameters, it helps with overfitting when dataset is small, so when risk is higher due to low quantity of data.<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Deep Learning: substitute a model with a neural network for analysis.<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Deep Learning features can be analysed as raw because in complex network, layers themselves are able to transform low level features in higher ones.<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">In Deep learning we have different types of <b>activation function<\/b>:<\/span><span class=\"s1\"><br \/>\n<\/span><\/p>\n<ol class=\"ol1\">\n<li class=\"li4\"><span class=\"s2\">Sigmoid<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">ReLU<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">Softplus<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<li class=\"li4\"><span class=\"s2\">TanH<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<\/ol>\n<\/li>\n<li class=\"li4\"><span class=\"s2\">When we talk about Deep Learning <b>CNN<\/b> or Convolutional Neural Network (very good for image processing) and <b>Long Short Term Memory,<\/b> are the most used for versatility and scalability.\u00a0<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<\/ol>\n<h2 class=\"p5\"><span class=\"s2\">Questions: where to find them?<\/span><\/h2>\n<p class=\"p12\"><span class=\"s10\">In general, because of AWS\u2019 NDA about exam\u2019s questions, its difficult to obtain <b>official<\/b> material (or at least <\/span><span class=\"s11\">officious and reliable)&#8230;This list of links points to resources personally selected by me as proven to be verified, trustable and functional to certification\u2019s preparation:<\/span><\/p>\n<ul class=\"ul1\">\n<li class=\"li13\"><b><\/b><span class=\"s10\"><b>Official AWS questions (free): <\/b><a href=\"https:\/\/d1.awsstatic.com\/training-and-certification\/docs-ml\/AWS-Certified-Machine-Learning-Specialty_Sample-Questions.pdf\"><span class=\"s9\">https:\/\/d1.awsstatic.com\/training-and-certification\/docs-ml\/AWS-Certified-Machine-Learning-Specialty_Sample-Questions.pdf<\/span><\/a><\/span><span class=\"s13\"><br \/>\n<\/span><\/li>\n<\/ul>\n<ul class=\"ul1\">\n<li class=\"li13\"><b><\/b><span class=\"s10\"><b>Official AWS test exam (paid): <\/b><a href=\"https:\/\/www.aws.training\/certification?rightcta=mlexam\"><span class=\"s9\">https:\/\/www.aws.training\/certification?rightcta=mlexam<\/span><\/a><\/span><span class=\"s13\"><br \/>\n<\/span><\/li>\n<\/ul>\n<ul class=\"ul1\">\n<li class=\"li13\"><b><\/b><span class=\"s10\"><b>Exam readiness class (paid): <\/b><a href=\"https:\/\/www.aws.training\/SessionSearch?pageNumber=1&amp;courseId=38153\"><span class=\"s9\">https:\/\/www.aws.training\/SessionSearch?pageNumber=1&amp;courseId=38153<\/span><\/a><\/span><span class=\"s13\"><br \/>\n<\/span><\/li>\n<\/ul>\n<ul class=\"ul1\">\n<li class=\"li4\"><b><\/b><span class=\"s2\"><b>Certbolt test exam with 65 questions<\/b> &#8211; only on Windows system or VM &#8211; <\/span><span class=\"s14\"><b>highly recommended (free)<\/b><\/span><span class=\"s2\">: <a href=\"https:\/\/www.certbolt.com\/aws-certified-machine-learning-specialty-exam-dumps\"><span class=\"s9\">https:\/\/www.certbolt.com\/aws-certified-machine-learning-specialty-exam-dumps<\/span><\/a><\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<\/ul>\n<ul class=\"ul1\">\n<li class=\"li13\"><b><\/b><span class=\"s10\"><b>Some studying materials from Udemy (paid)<\/b>: <a href=\"https:\/\/www.udemy.com\/topic\/aws-certified-machine-learning-specialty\/\"><span class=\"s9\">https:\/\/www.udemy.com\/topic\/aws-certified-machine-learning-specialty\/<\/span><\/a><\/span><span class=\"s13\"><br \/>\n<\/span><\/li>\n<\/ul>\n<ul class=\"ul1\">\n<li class=\"li4\"><span class=\"s2\">Various questions: found on Google by searching \u201c<b>aws machine learning specialty dumps<\/b>\u201d. They can be useful to verify your knowledge but <b>not recommended as many of them are unfortunately not correct<\/b>. Use them just as a proof of your own knowledge.\u00a0<\/span><span class=\"s1\"><br \/>\n<\/span><\/li>\n<\/ul>\n<h2 class=\"p14\"><span class=\"s1\">Side note: remote exam with Pearson\/VUE<\/span><\/h2>\n<p class=\"p12\"><span class=\"s11\">The remote exam involves accessing the Pearson\/VUE panel and registering for a date when online proctoring is available, necessary to ascertain the validity of the exam. Given the emergency situation this year, it is advisable to book the session <b>well in advance<\/b>, to be sure of finding a convenient date and time.<\/span><\/p>\n<p class=\"p12\"><span class=\"s11\">You must have a valid ID card or <b>driving license or passport<\/b>, of which clear and understandable photos will be required. A damaged document or an illegible photo will cause significant delays in the scheduling due to telephone inquiries.<\/span><\/p>\n<p class=\"p12\"><span class=\"s11\">Software available for all Windows, Mac and Linux platforms will be downloaded, which will be used to take the exam and which will verify the room in which the test will take place by means of photos from the mobile phone. Also in this case it is necessary to scrupulously follow the instructions received via email upon registration, to carry out the validation tests before the exam date.<\/span><\/p>\n<p class=\"p12\"><span class=\"s11\">During the examination, a proctor will follow you from your webcam and the software downloaded will run in fullscreen to prevent the use of other programs.<\/span><\/p>\n<p class=\"p12\"><span class=\"s11\">Should any problem occur, the proctor will notify you via internal chat.<\/span><\/p>\n<p class=\"p12\"><span class=\"s11\">To conclude, if you were wondering if it was possible to prepare the AWS Certified Machine Learning Specialty &#8211; or an AWS specialty in general &#8211; on your own, now you know that the answer is &#8230; <strong>technically yes<\/strong> \ud83d\ude42<\/span><\/p>\n<p class=\"p12\"><span class=\"s11\">However, for those who are beginners with AWS certifications or those who need structured training, it is important to know that <strong>AWS puts official courses at your disposal<\/strong> by delivering them through recognized APN Training Partners. <strong>Amazon Web Services has selected beSharp as one of the few APN Training Partners authorized to provide official AWS courses.<\/strong> Our training classes are developed and managed by AWS multi-certified Cloud Experts to ensure that content reflects the latest best practices. The classroom training offers participants the opportunity to receive live feedback and answers to questions directly from an expert instructor. They also include practical <strong>workshops<\/strong> designed to consolidate your knowledge.<\/span><\/p>\n<p class=\"p12\"><span class=\"s11\">If you want to prepare this or any other official AWS certification, <a href=\"https:\/\/www.besharp.it\/en\/aws-training\/\" target=\"_blank\" rel=\"noopener noreferrer\">read all the details<\/a> and <a href=\"https:\/\/www.besharp.it\/en\/contact-us\/\" target=\"_blank\" rel=\"noopener noreferrer\">contact us to plan your training.<\/a><\/span><\/p>\n<p class=\"p12\"><span class=\"s11\">Good luck with your next AWS certification and see you soon!<\/span><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Introduction He did it! Our Cloud Expert Alessandro Gaggia got his sixth (!) AWS Certification (the 58th AWS Certification for [&hellip;]<\/p>\n","protected":false},"author":6,"featured_media":1486,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[480],"tags":[349,368,324],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v21.9.1 - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>AWS Machine Learning Specialty: How I got certified in ten days as a Devops Engineer - Proud2beCloud Blog<\/title>\n<meta name=\"description\" content=\"AWS Machine Learning Specialty: How I got certified in ten days as a Devops Engineer\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/blog.besharp.it\/aws-machine-learning-specialty-how-i-got-certified-in-ten-days-as-a-devops-engineer\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"AWS Machine Learning Specialty: How I got certified in ten days as a Devops Engineer\" \/>\n<meta property=\"og:description\" content=\"How I got my AWS Machine Learning Specialty certfication as a Devops Engineer\" \/>\n<meta property=\"og:url\" content=\"https:\/\/blog.besharp.it\/aws-machine-learning-specialty-how-i-got-certified-in-ten-days-as-a-devops-engineer\/\" \/>\n<meta property=\"og:site_name\" content=\"Proud2beCloud Blog\" \/>\n<meta property=\"article:published_time\" content=\"2020-07-10T10:02:47+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2024-02-02T10:56:31+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/blog.besharp.it\/wp-content\/uploads\/2020\/07\/AWS_ML_Specialty_get_certified.png\" \/>\n\t<meta property=\"og:image:width\" content=\"834\" \/>\n\t<meta property=\"og:image:height\" content=\"626\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"author\" content=\"Alessandro Gaggia\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:title\" content=\"AWS Machine Learning Specialty: How I got certified in ten days as a Devops Engineer\" \/>\n<meta name=\"twitter:description\" content=\"How I got my AWS Machine Learning Specialty certfication as a Devops Engineer\" \/>\n<meta name=\"twitter:image\" content=\"https:\/\/blog.besharp.it\/wp-content\/uploads\/2020\/07\/AWS_ML_Specialty_get_certified.png\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Alessandro Gaggia\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"11 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/blog.besharp.it\/aws-machine-learning-specialty-how-i-got-certified-in-ten-days-as-a-devops-engineer\/\",\"url\":\"https:\/\/blog.besharp.it\/aws-machine-learning-specialty-how-i-got-certified-in-ten-days-as-a-devops-engineer\/\",\"name\":\"AWS Machine Learning Specialty: How I got certified in ten days as a Devops Engineer - Proud2beCloud Blog\",\"isPartOf\":{\"@id\":\"https:\/\/blog.besharp.it\/#website\"},\"datePublished\":\"2020-07-10T10:02:47+00:00\",\"dateModified\":\"2024-02-02T10:56:31+00:00\",\"author\":{\"@id\":\"https:\/\/blog.besharp.it\/#\/schema\/person\/f27fc12d10867c6ea6e0158ce4dd8924\"},\"description\":\"AWS Machine Learning Specialty: How I got certified in ten days as a Devops Engineer\",\"breadcrumb\":{\"@id\":\"https:\/\/blog.besharp.it\/aws-machine-learning-specialty-how-i-got-certified-in-ten-days-as-a-devops-engineer\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/blog.besharp.it\/aws-machine-learning-specialty-how-i-got-certified-in-ten-days-as-a-devops-engineer\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/blog.besharp.it\/aws-machine-learning-specialty-how-i-got-certified-in-ten-days-as-a-devops-engineer\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/blog.besharp.it\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"AWS Machine Learning Specialty: How I got certified in ten days as a Devops Engineer\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/blog.besharp.it\/#website\",\"url\":\"https:\/\/blog.besharp.it\/\",\"name\":\"Proud2beCloud Blog\",\"description\":\"il blog di beSharp\",\"alternateName\":\"Proud2beCloud Blog\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/blog.besharp.it\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"en-US\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/blog.besharp.it\/#\/schema\/person\/f27fc12d10867c6ea6e0158ce4dd8924\",\"name\":\"Alessandro Gaggia\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/blog.besharp.it\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/secure.gravatar.com\/avatar\/f58dc28050f26409e22ab60346d06220?s=96&d=mm&r=g\",\"contentUrl\":\"https:\/\/secure.gravatar.com\/avatar\/f58dc28050f26409e22ab60346d06220?s=96&d=mm&r=g\",\"caption\":\"Alessandro Gaggia\"},\"description\":\"Head of software development di beSharp, Full-Stack developer, mi occupo di garantire lo stato dell\u2019arte di tutta la nostra codebase. Scrivo codice in quasi ogni linguaggio, ma prediligo Typescript. Respiro Informatica, Game design, Cinema, Fumetti e buona cucina. Disegno per passione!\",\"url\":\"https:\/\/blog.besharp.it\/author\/alessandro-gaggia\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"AWS Machine Learning Specialty: How I got certified in ten days as a Devops Engineer - Proud2beCloud Blog","description":"AWS Machine Learning Specialty: How I got certified in ten days as a Devops Engineer","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/blog.besharp.it\/aws-machine-learning-specialty-how-i-got-certified-in-ten-days-as-a-devops-engineer\/","og_locale":"en_US","og_type":"article","og_title":"AWS Machine Learning Specialty: How I got certified in ten days as a Devops Engineer","og_description":"How I got my AWS Machine Learning Specialty certfication as a Devops Engineer","og_url":"https:\/\/blog.besharp.it\/aws-machine-learning-specialty-how-i-got-certified-in-ten-days-as-a-devops-engineer\/","og_site_name":"Proud2beCloud Blog","article_published_time":"2020-07-10T10:02:47+00:00","article_modified_time":"2024-02-02T10:56:31+00:00","og_image":[{"width":834,"height":626,"url":"https:\/\/blog.besharp.it\/wp-content\/uploads\/2020\/07\/AWS_ML_Specialty_get_certified.png","type":"image\/png"}],"author":"Alessandro Gaggia","twitter_card":"summary_large_image","twitter_title":"AWS Machine Learning Specialty: How I got certified in ten days as a Devops Engineer","twitter_description":"How I got my AWS Machine Learning Specialty certfication as a Devops Engineer","twitter_image":"https:\/\/blog.besharp.it\/wp-content\/uploads\/2020\/07\/AWS_ML_Specialty_get_certified.png","twitter_misc":{"Written by":"Alessandro Gaggia","Est. reading time":"11 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/blog.besharp.it\/aws-machine-learning-specialty-how-i-got-certified-in-ten-days-as-a-devops-engineer\/","url":"https:\/\/blog.besharp.it\/aws-machine-learning-specialty-how-i-got-certified-in-ten-days-as-a-devops-engineer\/","name":"AWS Machine Learning Specialty: How I got certified in ten days as a Devops Engineer - Proud2beCloud Blog","isPartOf":{"@id":"https:\/\/blog.besharp.it\/#website"},"datePublished":"2020-07-10T10:02:47+00:00","dateModified":"2024-02-02T10:56:31+00:00","author":{"@id":"https:\/\/blog.besharp.it\/#\/schema\/person\/f27fc12d10867c6ea6e0158ce4dd8924"},"description":"AWS Machine Learning Specialty: How I got certified in ten days as a Devops Engineer","breadcrumb":{"@id":"https:\/\/blog.besharp.it\/aws-machine-learning-specialty-how-i-got-certified-in-ten-days-as-a-devops-engineer\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/blog.besharp.it\/aws-machine-learning-specialty-how-i-got-certified-in-ten-days-as-a-devops-engineer\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/blog.besharp.it\/aws-machine-learning-specialty-how-i-got-certified-in-ten-days-as-a-devops-engineer\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/blog.besharp.it\/"},{"@type":"ListItem","position":2,"name":"AWS Machine Learning Specialty: How I got certified in ten days as a Devops Engineer"}]},{"@type":"WebSite","@id":"https:\/\/blog.besharp.it\/#website","url":"https:\/\/blog.besharp.it\/","name":"Proud2beCloud Blog","description":"il blog di beSharp","alternateName":"Proud2beCloud Blog","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/blog.besharp.it\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"en-US"},{"@type":"Person","@id":"https:\/\/blog.besharp.it\/#\/schema\/person\/f27fc12d10867c6ea6e0158ce4dd8924","name":"Alessandro Gaggia","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/blog.besharp.it\/#\/schema\/person\/image\/","url":"https:\/\/secure.gravatar.com\/avatar\/f58dc28050f26409e22ab60346d06220?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/f58dc28050f26409e22ab60346d06220?s=96&d=mm&r=g","caption":"Alessandro Gaggia"},"description":"Head of software development di beSharp, Full-Stack developer, mi occupo di garantire lo stato dell\u2019arte di tutta la nostra codebase. Scrivo codice in quasi ogni linguaggio, ma prediligo Typescript. Respiro Informatica, Game design, Cinema, Fumetti e buona cucina. Disegno per passione!","url":"https:\/\/blog.besharp.it\/author\/alessandro-gaggia\/"}]}},"_links":{"self":[{"href":"https:\/\/blog.besharp.it\/wp-json\/wp\/v2\/posts\/1487"}],"collection":[{"href":"https:\/\/blog.besharp.it\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blog.besharp.it\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blog.besharp.it\/wp-json\/wp\/v2\/users\/6"}],"replies":[{"embeddable":true,"href":"https:\/\/blog.besharp.it\/wp-json\/wp\/v2\/comments?post=1487"}],"version-history":[{"count":0,"href":"https:\/\/blog.besharp.it\/wp-json\/wp\/v2\/posts\/1487\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/blog.besharp.it\/wp-json\/wp\/v2\/media\/1486"}],"wp:attachment":[{"href":"https:\/\/blog.besharp.it\/wp-json\/wp\/v2\/media?parent=1487"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blog.besharp.it\/wp-json\/wp\/v2\/categories?post=1487"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blog.besharp.it\/wp-json\/wp\/v2\/tags?post=1487"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}