Progress continues on our upcoming Machine Learning Mastery course. This course shows you how to process data in real-time from APIs and how to deploy artificial intelligence models into production.

As an introduction to the course, I invite you to watch the following video below which goes over the basic roadmap of how to run your own machine learning models in real-time (the very cool motion graphics and step-by-step flowchart are roughly 6 minutes into the video).

Finally, to give you a sneak preview of what we are putting together, take a look at the outline for the course. As you can see, the course is going to be quite comprehensive in order to give students all the tools that are needed to build out a basic machine learning pipeline.

Machine Learning Mastery Course Outline

1 - General Overview and Setup

  • General Overview
  • Installing Linux on Windows
  • Downloading Full Code and Overview of Codebase
  • Installing Python
  • Installing PyCharm
  • Installation of python packages for the course
  • Installing Docker Desktop
  • Installation of command line utilities for the cloud

2 - Streaming Data with Docker

  • Why use an isolated environment to pull in Twitter data?
  • Tweet Filter Overview
  • Configure your python program
  • Using requests to stream data
  • Pub/Sub overview
  • Publishing to a topic
  • Dockerfiles 101
  • Testing Locally
  • Deployment to Google's Container Registry
  • Running the Container in the cloud
  • Cloud Logging Overview

3 - Apache Beam Basics

  • What are we doing?
  • Beam - Group by Key
  • Inspecting PCollections
  • Counting Big Data
  • Filtering
  • Using Failsafe Elements in the data pipeline
  • Filter non-english tweets

4 - BigQuery

  • BigQuery Overview
  • SQL 101
  • Building pipeline export to Apache Beam
  • Deploying the data pipeline
  • Analyze Pipeline results

5 - Creating a Machine Learning Model

  • ML Overview
  • Develop models with free GPU on Google Colab
  • Ingesting Labeled Data
  • Importance of Tokenization
  • Keras 101
  • ML Concepts: CNN Model Overview
  • ML Concepts: MaxPool
  • ML Concepts: Dropout
  • Training a machine learning model
  • Testing accuracy of the machine learning model

6 - Model Deployment

  • Strengths and Weaknesses of AI Platform
  • Model packaging
  • Model Deployment
  • Calling the model as an API from Python

7 - Incorporating Model into the Production Pipeline

  • Adding a validator to verify parameters
  • Time Windows in data pipelines
  • Calling the Machine Learning API from Apache Beam
  • Routing the data pipeline to BigQuery
  • Verifying model prediction results with SQL in BigQuery

8 - Visualizing Your Results

  • Data Studio Overview
  • Ingesting BigQuery data
  • Showing results in charts

If you like what you see, follow White Owl Education on Twitter or on Facebook for future updates.