Terminology Breakdown
AI
Machine Learning
Deep Learning
Training & Inference

Determining if the problem is suitable for machine learning

Types of Machine Learning and their use case
Regression
Classification
Computer Vision
•   Image classification
•   Object detection
•   Semantic segmentation

Why we need a machine learning pipeline

Training a Model – Exploratory Data Analysis
Getting to know your data
Using low-code tools to visualize and clean data
Using Python libraries to plot charts

Training a Model – Preparing Your Data for Machine Learning Training
Dealing with missing data
Finding invalid data
Understanding your features and their correlation
Measuring data distribution and determining what can be done
Handling outliers
Scaling your numeric features
Why we need feature engineering
One-hot encoding for categorical data

Training a Model – Choosing an algorithm
Why would you choose a particular algorithm such as XGBoost, LinearLearner or K-means

Training & evaluating a model What are Hyper-parameters
What are hold out data sets and why we need them in evaluation
Understanding the objective loss function
Interpreting a confusion matrix

Inference – Hosting a model
Hosting your trained model in Amazon SageMaker

Inference – Running inference and obtaining prediction
Creating an inference batch request
Createing an inference real-time request

Inference – Monitoring a model
Understanding and detecting data drift
Understanding and remediating concept drift