Ingénieur en apprentissage automatique AWS SageMaker en 30 jours + ChatGPT

Aperçu

Build 30+ ML Projects in 30 Days in AWS, Master SageMaker JumpStart, Canvas, AutoPilot, DataWrangler, Lambda & S3 What you'll learn:

Build, Train, Test and Deploy Machine Learning Models in AWSLeverage ChatGPT and GPT-4 to Automate Coding Tasks, Perform Code Debugging, Write Documentation and Add New Features to your CodeDefine and Perform Image and Text Labeling Jobs Using AWS SageMaker GroundTruthPrepare, Clean and Visualize data Using AWS SageMaker Data Wrangler without Writing any CodeOptimize ML model hyperparameters using GridSearch, Bayesian & Random Search Optimization TechniquesMaster Key AWS services such as Simple Storage Service (S3), Elastic Compute Cloud (EC2), Identity and Access Management (IAM) and CloudWatchUnderstand Machine Learning workflow automation using AWS Lambda, Step functions and SageMaker Pipelines.Learn how to define a lambda function in AWS management console, understand the anatomy of Lambda functions, and how to configure a test event in LambdaTrain a Machine Learning Regression and Classifier Models Using No-code AWS CanvasLearn how to leverage Amazon SageMaker Autopilot and SageMaker Canvas to train multiple models without writing any code.Perform Exploratory Data Analysis and Visualization Using Pandas, Searborn and Matplotlib LibrariesUnderstand Regression Models KPIs Such as RMSE, MSE, MAE, R2 and Adjusted R2Understand Classification Models KPIs such as Accuracy, Precision, Recall, F1-Score, ROC, and AUCDefine a Machine Learning Training Job Using AWS SageMaker JumpStartDeploy an Endpoint Using Amazon SageMaker, Perform Inference and Generate PredictionsDefine a Lambda function using Boto3 SDK and Test the lambda function using Eventbridge (cloudwatch events)Understand the difference between synchronous and asynchronous Lambda Functions invocationsPerform AI/ML Models Prototyping Using AutoGluon LibraryHow to monitor billing dashboard, set alarms, S3/EC2 instances pricing and request service limits increaseUnderstand the difference between Artificial Intelligence (AI), Machine Learning (ML), Data Science (DS) and Deep Learning (DL)Learn the fundamentals of Amazon SageMaker, SageMaker Components, training options including built-in algorithms, AWS Marketplace, & customized ML AlgorithmsLeverage a Yolo V3 Object Detection Algorithm available on the AWS MarketplaceUnderstand the format and Use Case of Json Lines and Manifest FilesLearn auto-labeling workflow and understand the difference between SageMaker GroundTruth and GroundTruth PlusLearn how to define a labeling job with bounding boxes (object detection), pixel-level Semantic Segmentation, and text dataUnderstand the difference between data labeling workforces in AWS such as public mechanical Turks, private labelers and AWS curated third-party vendorsLearn the difference between Supervised, Unsupervised and Reinforcement Machine Learning StrategiesPerform data visualization using Seaborn & Matplotlib libraries, plots include line plot, pie charts, subplots, pairplots, countplots, and correlations heatmapsExport a data wrangler workflow into Python script, create a custom formula and apply it to a given column in the data, and generate summary tables/bias reportLearn how to train an XG-boost algorithm in SageMaker using AWS JumpStart, assess trained model performance, plot residuals, & deploy an endpointUnderstand Bias-Variance Trade-off, L1 and L2 Regularization TechniquesTrain/Test several ML Classifiers such as Logistic Regression, Support Vector Machine, K-Nearest Neighbors, Decision Trees, and Random Forest ClassifiersLearn SageMaker Built-in Algorithms such as Linear Learner, XG-Boost, Principal Component Analysis (PCA), and K-Nearest Neighbors Do you want to become an AWS Machine Learning Engineer Using SageMaker in 30 days?Do you want to build super-powerful production-level Machine Learning (ML) applications in AWS but don’t know where to start?Are you an absolute beginner and want to break into AI, ML, and Cloud Computing and looking for a course that includes everything you need?Are you an aspiring entrepreneur who wants to maximize business revenues and reduce costs with ML but don’t know how to get there quickly and efficiently?Do you want to leverage ChatGPTas a programmer to automate your coding tasks?If the answer is yes to any of these questions, then this course is for you!Machine Learning is the future one of the top tech fields to be in right now! ML and AI will change our lives in the same way electricity did 100 years ago.

ML is widely adopted in Finance, banking, healthcare, transportation, and technology. The field is exploding with opportunities and career prospectsAWS is the one of the most widely used cloud computing platforms in the world and several companies depend on AWS for their cloud computing purposes.

AWS SageMaker is a fully managed service offered by AWS that allows data scientist and AI practitioners to train, test, and deploy AI/ML models quickly and efficiently.This course is unique and exceptional in many ways, it includes several practice opportunities, quizzes, and final capstone projects. In this course, students will learn how to create production-level ML models using AWS.

The course is divided into 8 main sections as follows:

Section 1 (Days 1 – 3):

we will learn the following:

(1) Start with an AWS and Machine Learning essentials “starter pack” that includes key AWS services such as Simple Storage Service (S3), Elastic Compute Cloud (EC2), Identity and Access Management (IAM) and CloudWatch, (2) The benefits of cloud computing, the difference between regions and availability zones and what’s included in the AWS Free Tier Package, (3) How to setup a brand-new account in AWS, setup a Multi-Factor Authentication (MFA) and navigate through the AWS Management Console, (4) How to monitor billing dashboard, set alarms, S3/EC2 instances pricing and request service limits increase, (5) The fundamentals of Machine Learning and understand the difference between Artificial Intelligence (AI), Machine Learning (ML), Data Science (DS) and Deep Learning (DL), (6) Learn the difference between supervised, unsupervised and reinforcement learning, (7) List the key components to build any machine learning models including data, model, and compute, (8) Learn the fundamentals of Amazon SageMaker, SageMaker Components, training options offered by SageMaker including built-in algorithms, AWS Marketplace, and customized ML algorithms, (9) Cover AWS SageMaker Studio and learn the difference between AWS SageMaker JumpStart, SageMaker Autopilot and SageMaker Data Wrangler, (10) Learn how to write our first code in the cloud using Jupyter Notebooks. We will then have a tutorial covering AWS Marketplace object detection algorithms such as Yolo V3, (11) Learn how to train our first machine learning model using the brand-new AWS SageMaker Canvas without writing any code!Section 2 (Days 4 – 5):

we will learn the following:

(1) Label images and text using Amazon SageMaker GroundTruth, (2) learn the difference between data labeling workforces such as public mechanical Turks, private labelers and AWS curated third-party vendors, (3) cover several companies’ success stories that have leveraged data to maximize revenues, reduce costs and optimize processes, (4) cover data sources, types, and the difference between good and bad data, (5) learn about Json Lines formats and Manifest Files, (6) cover a detailed tutorial to define an image classification labeling job in SageMaker, (7) auto-labeling workflow and learn the difference between SageMaker GroundTruth and GroundTruth Plus, (8) learn how to define a labeling job with bounding boxes (object detection and pixel-level Semantic Segmentation), (9) Label Text data using Amazon SageMaker GroundTruth.Section 3 (Days 6 – 10):

we will learn:

(1) how to perform exploratory data analysis (EDA), (2) master Pandas, a super powerful open-source library to perform data analysis in Python, (3) analyze corporate employee information using Pandas in Jupyter Notebooks in AWS SageMaker Studio, (4) define a Pandas Dataframe, read CSV data using Pandas, perform basic statistical analysis on the data, set/reset Pandas DataFrame index, select specific columns from the DataFrame, add/delete columns from the DataFrame, Perform Label/integer-based elements selection, perform broadcasting operations, and perform Pandas DataFrame sorting/ordering, (5) perform statistical data analysis on real world datasets, deal with missing data using pandas, change pandas DataFrame datatypes, define a function, and apply it to a Pandas DataFrame column, perform Pandas operations, and filtering, calculate and display correlation matrix, use seaborn library to show heatmap, (6) analyze cryptocurrency prices and daily returns of Bitcoin (BTC), Ethereum (ETH), Litecoin (LTC), Cardano (ADA) and Ripple (XRP) using Matplotlib and Seaborn libraries in AWS SageMaker Studio, (7) perform data visualization using Seaborn and Matplotlib libraries, plots include line plot, pie charts, multiple subplots, pairplot, count plot, correlations heatmaps, distribution plot (distplot), Histograms, and Scatterplots, (8) Use Amazon SageMaker Data wrangler in AWS to prepare, clean and visualize the data, (9) understand feature engineering strategies and tools, understand the fundamentals of Data Wrangler in AWS, perform one hot encoding and normalization, perform data visualization Using Data Wrangler, export a data wrangler workflow into Python script, create a custom formula and apply it to a given column in the data, generate summary table tables in Data Wrangler, and generate bias reports.Section 4 (Days 11 – 18):

we will learn:

(1) machine learning regression fundamentals including simple/multiple linear regression and least sum of squares, (2) build our first simple linear regression model in Scikit-Learn, (3) list all available built-in algorithms in SageMaker, (4) build, train, test and deploy a machine learning regression model using SageMaker Linear Learner algorithm, (5) list machine learning regression algorithms KPIs such as Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Percentage Error (MPE), Coefficient of Determination (R2), and adjusted R2, (6) Launch a training job using the AWS Management Console and deploy an endpoint without writing any code, (7) cover the theory and intuition behind XG-Boost algorithm and how to use it to solve regression type problems in Scikit-Learn and using SageMaker Built-in algorithms, (8) learn how to train an XG-boost algorithm in SageMaker using AWS JumpStart, assess trained regression models performance, plot the residuals, and deploy an endpoint and perform inference.Section 5 (Days 19 – 20):

we will learn:

(1) hyperparameters optimization strategies such as grid search, randomized search, and Bayesian optimization, (2) Understand bias variance trade-off and L1 and L2 regularization, (3) perform hyperparameters optimization using Scikit-Learn library and using SageMaker SDK.Section 6 (Days 21 – 24):

we will learn:

(1) how to train several classification algorithms such as Logistic Regression, Support Vector Machine, K-Nearest Neighbors, and Random Forest Classifier, (2) list the difference between various classifier models KPIs such as accuracy, precision, recall, F1-score, Receiver Operating Characteristic Curve (ROC) and Area Under the Curve (AUC), (3) train an XG-boost and Linear Learner algorithms in SageMaker to solve classification type problems, (4) learn the theory and intuition behind K Nearest Neighbors (KNN) in SageMaker and learn how to build, train and test a KNN classifier model in SageMaker.

This section also includes bonus materials on how to leverage ChatGPT and generative AImodels as a programmer. Section 7 (Days 25 – 28):

we will learn:

(1) how to use AutoGluon library to perform prototyping of AI/ML models using few lines of code, (2) leverage AutoGluon to train multiple regression and classification models and deploy the best one, (3) leverage Amazon SageMaker Autopilot and SageMaker Canvas to train multiple models without writing any code.Section 8 (Days 29 – 30):

we will learn:

(1) how to define and invoke lambda functions in AWS, (2) understand Machine Learning workflow automation using AWS Lambda, Step functions and SageMaker Pipelines, (3) learn how to define a lambda function in AWS management console, (4) understand the anatomy of Lambda functions, (5) learn how to configure a test event in Lambda, and monitor Lambda invocations in CloudWatch, (6) define a Lambda function using Boto3 SDK, (7) test the lambda function using Eventbridge (cloudwatch events), (8) understand the difference between synchronous and asynchronous invocations, and Invoke a Lambda function using Boto3 SDK.

Programme

• Introduction à AWS SageMaker
Aperçu de AWS SageMaker et de ses composants
Configuration du compte AWS et de l'environnement SageMaker
Navigation dans la console SageMaker
• Fondamentaux de l'apprentissage automatique
Concepts clés de l'apprentissage automatique
Types d'apprentissage automatique : supervisé, non supervisé, par renforcement
Préparation et prétraitement des données
• SageMaker Studio et Jupyter Notebooks
Introduction à SageMaker Studio
Utilisation de Jupyter Notebooks pour les workflows de ML
Exploration des jeux de données intégrés et externes
• Ingénierie des données dans SageMaker
Solutions de stockage de données : S3, Redshift et RDS
Transformation et nettoyage des données avec SageMaker Data Wrangler
Meilleures pratiques en ingénierie des fonctionnalités
• Construction et entraînement de modèles
Sélection et utilisation des algorithmes intégrés
Développement de modèles personnalisés avec Python et TensorFlow/PyTorch
Gestion de l'entraînement de modèles à grande échelle : réglage des hyperparamètres et entraînement distribué
• Évaluation et déploiement de modèles
Évaluation des performances des modèles avec SageMaker
Test A/B et validation de modèles
Déploiement de modèles à l'aide de SageMaker Endpoints
• Automatisation et pipelines
Création de workflows ML automatisés avec SageMaker Pipelines
Utilisation d'Amazon Step Functions pour orchestrer les workflows ML
Surveillance et gestion des exécutions de pipeline
• Gestion et optimisation des coûts
Comprendre la tarification de SageMaker
Stratégies pour une utilisation rentable des ressources AWS
Surveillance de l'utilisation et réduction des coûts
• Sujets avancés de SageMaker
SageMaker RL pour les projets d'apprentissage par renforcement
Exploration de SageMaker JumpStart pour des solutions préconçues
Intégration avec d'autres services AWS comme Lambda et API Gateway
• Introduction à ChatGPT
Aperçu de ChatGPT et de ses capacités
Création d'agents conversationnels avec ChatGPT
Intégration de ChatGPT dans des applications avec Amazon SageMaker
• Projet de fin d'études
Conception d'une solution ML réelle en utilisant SageMaker
Mise en œuvre d'un chatbot en utilisant ChatGPT
Présentation et évaluation des résultats du projet
• Revue finale et évaluation
Revue complète des sujets clés
Examens pratiques et tests de simulation
Conseils et ressources pour la préparation à la certification

---

Enseigné par

Dr. Ryan Ahmed, Ph.D., MBA

---

Matières

Programming