Deploying Data Science Models on GCP Online Course

Deploying Data Science Models on GCP Online Course

This course is designed for aspiring cloud engineers and data scientists, covering Google Cloud Platform's key serverless components. Learn to implement machine learning pipelines using Vertex AI with Kubeflow and manage Serverless PySpark through Dataproc, App Engine, and Cloud Run. The course starts with cloud concepts, setting up GCP accounts, and using tools like Cloud Functions, Cloud Run, and Vertex AI for custom model development. You'll also dive into serverless applications, event-driven architectures, and orchestration with Kubeflow. By the end, you'll be equipped to deploy and scale applications using GCP’s serverless components and Spark.

Who is this book for?

This intermediate course is aimed at aspiring data scientists, machine learning engineers, data engineers, architects, and IT professionals looking to transition into cloud technologies. It's ideal for individuals with a basic understanding of cloud concepts and experience in Python, SQL, and the Bash command line. A foundational tech background and familiarity with programming will help accelerate learning and grasp cloud concepts more effectively.

What you will learn

• Deploy serverless apps with Google App Engine, Cloud Functions, and Cloud Run
• Use datastore (NoSQL database) for real-world applications
• Explore microservice and event-driven architecture through practical examples
• Deploy production-level ML workflows on the cloud
• Orchestrate machine learning with Kubeflow using Python
• Deploy and schedule Serverless PySpark Jobs with Dataproc Serverless and Airflow/Composer

Course Table of Contents

Course Introduction and Prerequisites

• Course Introduction and Section Walkthrough
• Course Prerequisites

Modern-Day Cloud Concepts

• Introduction
• Scalability - Horizontal Versus Vertical Scaling
• Serverless Versus Servers and Containerization
• Microservice Architecture
• Event-Driven Architecture

Get Started with Google Cloud

• Set Up GCP Trial Account
• Google Cloud CLI Setup
• Get Comfortable with Basics of gcloud CLI
• gsutil and Bash Command Basics

Cloud Run - Serverless and Containerized Applications

• Section Introduction
• Introduction to Dockers
• Lab - Install Docker Engine
• Lab - Run Docker Locally
• Lab - Run and Ship Applications Using the Container Registry
• Introduction to Cloud Run
• Lab - Deploy Python Application to Cloud Run
• Cloud Run Application Scalability Parameters
• Introduction to Cloud Build
• Lab - Python Application Deployment Using Cloud Build
• Lab - Continuous Deployment Using Cloud Build and GitHub

Google App Engine - For Serverless Applications

• Introduction to App Engine
• App Engine - Different Environments
• Lab - Deploy Python Application to App Engine - Part 1
• Lab - Deploy Python Application to App Engine - Part 2
• Lab - Traffic Splitting in App Engine
• Lab - Deploy Python - BigQuery Application
• Caching and Its Use Cases
• Lab - Implement Caching Mechanism in Python Application - Part 1
• Lab - Implement Caching Mechanism in Python Application - Part 2
• Lab - Assignment Implement Caching
• Lab - Python App Deployment in a Flexible Environment
• Lab - Scalability and Instance Types in App Engine

Cloud Functions - Serverless and Event-Driven Applications

• Introduction
• Lab - Deploy Python Application Using Cloud Storage Triggers
• Lab - Deploy Python Application Using Pub/Sub Triggers
• Lab - Deploy Python Application Using HTTP Triggers
• Introduction to Cloud Datastore
• Overview Product Wishlist Use Case
• Lab – Use Case Deployment - Part-1
• Lab – Use Case Deployment - Part-2

Data Science Models with Google App Engine

• Introduction to ML Model Lifecycle
• Overview - Problem Statement
• Lab - Deploy Training Code to App Engine
• Lab - Deploy Model Serving Code to App Engine
• Overview - New Use Case
• Lab - Data Validation Using App Engine
• Lab - Workflow Template Introduction
• Lab - Final Solution Deployment Using Workflow and App Engine

Dataproc Serverless PySpark

• Introduction
• PySpark Serverless Autoscaling Properties
• Persistent History Cluster
• Lab - Develop and Submit PySpark Job
• Lab - Monitoring and Spark UI
• Introduction to Airflow
• Lab - Airflow with Serverless PySpark
• Wrap Up

Vertex AI - Machine Learning Framework

• Introduction
• Overview – Vertex AI UI
• Lab - Custom Model Training Using Web Console
• Lab - Custom Model Training Using SDK and Model Registries
• Lab - Model Endpoint Deployment
• Lab - Model Training Flow Using Python SDK
• Lab - Model Deployment Flow Using Python SDK
• Lab - Model Serving Using Endpoint with Python SDK
• Introduction to Kubeflow
• Lab - Code Walkthrough Using Kubeflow and Python
• Lab - Pipeline Execution in Kubeflow
• Lab - Final Pipeline Visualization Using Vertex UI and Walkthrough
• Lab - Add Model Evaluation Step in Kubeflow before Deployment
• Lab - Reusing Configuration Files for Pipeline Execution and Training
• Lab - Assignment Use Case - Fetch Data from BigQuery
• Wrap Up

Cloud Scheduler and Application Monitoring

• Introduction to Cloud Scheduler
• Lab - Cloud Scheduler in Action
• Lab - Set Up Alerting for Google App Engine Applications
• Lab - Set Up Alerting for Cloud-Run Applications
• Lab Assignment - Set Up Alerting for Cloud Function Applications