Machine Learning Engineer

This one-day instructor-led course introduces participants to the big data capabilities of Google Cloud Platform. Through a mix of presentations, demos, and hands-on labs, participants gain an understanding of the Google Cloud Platform and its data processing and machine learning capabilities. The course highlights the ease, flexibility, and power of big data solutions available on Google Cloud Platform.

Target Audience

• Data analysts, data scientists, and business analysts new to Google Cloud Platform.

• Professionals responsible for designing data processing pipelines, maintaining machine learning and statistical models, querying datasets, and creating reports.

• Executives and IT decision-makers evaluating Google Cloud Platform for data science use.

Prerequisites:

Participants should have:

• Basic proficiency in SQL or a similar query language.

• Experience with data modeling and ETL (Extract, Transform, Load) activities.

• Programming experience in languages such as Python.

• Familiarity with machine learning and/or statistics.

Learning Outcomes:

Participants will learn to:

• Identify the purpose and value of key Big Data and Machine Learning products in Google Cloud Platform.

• Use Cloud SQL and Cloud Dataproc to migrate MySQL and Hadoop/Pig/Spark/Hive workloads to Google Cloud Platform.

• Perform interactive data analysis with BigQuery and Cloud Datalab.

• Train and use neural networks with TensorFlow.

• Utilize ML APIs.

• Choose the appropriate data processing products on Google Cloud Platform.

Course Outline

Module 1: Google Cloud Big Data and Machine Learning Fundamentals
Topics:

• Big Data and Machine Learning on Google Cloud

• Data Engineering for Streaming Data

• Big Data with BigQuery

• Machine Learning Options on Google Cloud

• The Machine Learning Workflow with Vertex AI

Hands-On:

• AI Platform: Qwik Start

• Dataprep: Qwik Start

• Dataflow: Qwik Start - Templates

• Dataflow: Qwik Start - Python

• Dataproc: Qwik Start - Console

• Dataproc: Qwik Start - Command Line

Module 2: How Google Does Machine Learning
Topics:

• What It Means to be AI-First

• How Google Does ML

• Machine Learning Development with Vertex AI

• Machine Learning Development with Vertex Notebooks

• Best Practices for Implementing Machine Learning on Vertex AI

• Responsible AI Development

Hands-On:

• Vertex AI: Qwik Start

• Using an Image Dataset to Train an AutoML Model

Module 3: Launching into Machine Learning
Topics:

• Introduction

• Get to Know Your Data: Improve Data through Exploratory Data Analysis

• Machine Learning in Practice

• Training AutoML Models Using Vertex AI

• BigQuery Machine Learning: Develop ML Models Where Your Data Lives

• Optimization

• Generalization and Sampling

Hands-On:

• Exploratory Data Analysis Using Python and BigQuery

• Using BigQuery ML to Predict Penguin Weight

Module 4: TensorFlow on Google Cloud
Topics:

• Introduction to the TensorFlow ecosystem

• Design and Build an Input Data Pipeline

• Building Neural Networks with the TensorFlow and Keras API

• Training at Scale with Vertex AI

Hands-On:

• Classifying Structured Data using Keras Preprocessing Layers

• Build a DNN using the Keras Functional API

Module 5: Feature Engineering
Topics:

• Introduction to Vertex AI Feature Store

• Raw Data to Features

• Feature Engineering

• Preprocessing and Feature Creation

• Feature Crosses - TensorFlow Playground

• Introduction to TensorFlow Transform

Hands-On:

• Using Feature Store

• Performing Basic Feature Engineering in BQML

• Performing Basic Feature Engineering in Keras

Module 6: Machine Learning in the Enterprise
Topics:

• Introduction

• Understanding the ML Enterprise Workflow

• Data in the Enterprise

• Science of Machine Learning and Custom Training

• Vertex Vizier Hyperparameter Tuning

• Prediction and Model Monitoring Using Vertex AI

• Vertex AI Pipelines

• Best Practices for ML Development

Hands-On:

• Vertex Pipelines: Qwik Start

• Cloud Natural Language API: Qwik Start

• Google Cloud Speech API: Qwik Start

• Video Intelligence: Qwik Start

Module 7: End-to-End Machine Learning with TensorFlow on Google Cloud
Topics:

• Machine Learning (ML) on Google Cloud Platform (GCP)

• Explore the Data

• Create the dataset

• Build the Model

• Operationalize the model

Hands-On:

• Identify Damaged Car Parts with Vertex AutoML Vision

• Deploy a BigQuery ML Customer Churn Classifier to Vertex AI for Online Predictions

Module 8: Production Machine Learning Systems
Topics:

• Introduction to Advanced Machine Learning on Google Cloud

• Architecting Production ML Systems

• Designing Adaptable ML Systems

• Designing High-Performance ML Systems

• Building Hybrid ML Systems

Hands-On:

• Structured data prediction using Vertex AI Platform

• Serving ML Predictions in Batch and Real Time

• Distributed Training with Keras

• Using Kubeflow Pipelines with AI Platform

Module 9: Computer Vision Fundamentals with Google Cloud
Topics:

• Introduction to Computer Vision and Pre-built ML Models for Image Classification

• Vertex AI and AutoML Vision on Vertex AI

• Custom Training with Linear, Neural Network and Deep Neural Network models

• Convolutional Neural Networks

• Dealing with Image Data

Hands-On:

• Using the What-If Tool with Image Recognition Models

• Identifying Bias in Mortgage Data using Cloud AI Platform and the What-if Tool

• Compare Cloud AI Platform Models using the What-If Tool to Identify potential bias

Module 10: Sequence Models for Time Series and Natural Language Processing on Google Cloud
Topics:

• Working with Sequences

• Recurrent Neural Networks

• Dealing with Longer Sequences

• Text Classification

• Reusable Embeddings

• Encoder-Decoder Models

Hands-On:

• Time Series Prediction with a DNN Model

• Time Series Prediction with a Two-Layer RNN Model

• Text Classification using TensorFlow/Keras on AI Platform

• Text generation using tensor2tensor on Cloud AI Platform

Module 11: Recommendation Systems on Google Cloud
Topics:

• Recommendation Systems Overview

• Content-Based Recommendation Systems

• Collaborative Filtering Recommendations Systems

• Neural Networks for Recommendation Systems

• Reinforcement Learning

Hands-On:

• Using Neural Networks for Content-Based Filtering

• Collaborative Filtering on Google Analytics data

• ML on GCP: Hybrid Recommendations with the MovieLens Dataset

• Applying Contextual Bandits for Recommendations with Tensorflow and TFAgents

Module 12: MLOps (Machine Learning Operations) Fundamentals
Topics:

• Why and When do we Need MLOps

• Understanding the Main Kubernetes Components (Optional)

• Introduction to AI Platform Pipelines

• Training, Tuning and Serving on AI Platform

• Kubeflow Pipelines on AI Platform

• CI/CD for Kubeflow Pipelines on AI Platform

Hands-On:

• Working with Cloud Build

• Creating Google Kubernetes Engine Deployments

• Using custom containers with AI Platform Training

• Continuous Training Pipeline with Kubeflow Pipeline and Cloud AI Platform

• CI/CD for a Kubeflow pipeline

Module 13: ML Pipelines on Google Cloud
Topics:

• Introduction to TFX Pipelines

• Pipeline orchestration with TFX

• Custom components and CI/CD for TFX pipelines

• ML Metadata with TFX

• Continuous Training with multiple SDKs, KubeFlow & AI Platform Pipelines

• Continuous Training with Cloud Composer

• ML Pipelines with MLflow

Hands-On:

• TFX Standard Components Walkthrough

• TFX on Cloud AI Platform Pipelines

• CI/CD for a TFX pipeline

• Continuous Training Pipelines with Cloud Composer