AI 101 - Building Supervised ML Models

Welcome to this hands-on machine learning lab!

In this lab you will

• Work with real-world email spam data
• Build your own Transformer-based NLP model
• Train a classifier using deep learning
• Evaluate the performance of your model
• Understand each step of the ML pipeline
• Run live inference on user-provided text inputs

What This Lab Covers—and What It Does Not

• This lab focuses on supervised machine learning.
• The model produces deterministic predictions.
• The objective is classification accuracy, not text generation.
• Generative models and LLMs are covered in a separate advanced lab.

Learning Goals

By the end of this lab you will be able to:

• Explain how text datasets are prepared for machine learning
• Vectorize and tokenize text inputs
• Implement a Transformer encoder block
• Train a binary classification model
• Understand accuracy, AUC, confusion matrices
• Save and reload ML models
• Use a model for real-world predictions