AWS-Machine-Learning-Engineer-Nanodegree

ML Lifecycle, AWS Sagemaker, Autogluon, AWS Lambda, feature store, AWS step functions, Model Deployment, Deep Learning with CV and NLP, cloud performance management, Autoscaling, VPC

Certification Link

Introduction

This repository contains 5 Machine Learning Engineer projects using AWS Sagemaker and insights during my pursuit of the AWS MLE Nanodegree at Udacity:

Part 1. ML solutions in real world scenarios & AWS Sagemaker

• Feature engineering, ML use cases, Data loading with Sagemaker, AWS S3, XGBoost • Autogluon, Hyperparameter tuning, Dataset annotation, Sagemaker jumpstart, Automated ML, Sagemaker data wrangler
• Project: Predict Bike Sharing Demand with AutoGluon and AWS Sagemaker
• Report: Predict Bike Sharing Demand with AutoGluon Solution

Part 2. Developing ML Workflow & AWS Lambda & Step function & Model Endpoints

• AWS lambda, Sagemaker processing, Sagemaker batch transform jobs, Sagemaker training jobs, Sagemaker clarify, Machine learning pipeline creation, Sagemaker pipelines, Model monitoring, Sagemaker model endpoints, AWS step functions, Sagemaker model monitor, Sagemaker feature store
• Project: Build-ML-Workflow-Sagemaker-for-a-Logistics-Company

Part 3. Deep Learning Topics with Computer Vision and NLP & Sagemaker Profiler & Debugger

• Neural network basics, Deep learning fluency, Sagemaker jumpstart, Machine learning framework fundamentals, Hyperparameter tuning, Sagemaker training jobs, Transformer neural networks, Sagemaker debugger, Image classification, Training neural networks, Deep learning model optimization, Transfer learning, PyTorch, Model deployment with Sagemaker, Convolutional neural networks, Text classification, Model performance metrics, ETL pipeline, NLP pipeline, Machine Learning Pipeline, SQLite, Flask Web App
• Project: Image Classification using Sagemaker profiler and pre-trained model

Part 4. Operationalize ML to Production using SageMaker & Lambda & Feature store & Autoscaling

• Cloud resource allocation, AWS lambda, Distributed model training with Sagemaker, Amazon Elastic Compute Cloud, VPC, Sagemaker feature store, Cloud security in AWS, Cloud cost management, Sagemaker logs, Cloud performance management, AWS storage services, Training data manifest files, Sagemaker autoscaling
• Project: Operationalize ML pipeline to Production Server

Capstone Project. Build an ML end-to-end solution for Starbucks Marketing Strategies

• Customer segmentation, Behavioral Analysis, Predictive Model, Sagemaker SDK Deployment, AutoML, Endpoint, Model Querying, HTTP request
• Project: Starbucks Promotion Analysis and ML model Deployment & Testing