Build, Train and Deploy ML Models with Keras on Google Cloud Course Syllabus

Overview: This course offers a hands-on introduction to building, training, and deploying machine learning models using Keras and TensorFlow on Google Cloud. Designed for developers and data scientists, it blends foundational concepts with practical labs in Colab to reinforce real-world workflows. Through four structured modules, you'll progress from basic neural networks to advanced architectures, gaining experience with model optimization, persistence, and deployment. The course includes approximately 23 hours of content, featuring guided programming exercises and best practices from industry experts at DeepLearning.AI and Google. Ideal as a foundation for the TensorFlow Developer Professional Certificate.

Module 1: A New Programming Paradigm

Estimated time: 5 hours

• Introduction to machine learning and deep learning
• Understanding TensorFlow's programming paradigm
• Neural network basics and architecture
• Building a simple "Hello, World" neural network in Python

Module 2: The Sequential Model API

Estimated time: 6 hours

• Using Keras Sequential API for model construction
• Adding layers and compiling models
• Training and evaluating neural networks
• Building CNNs for MNIST digit classification in Colab

Module 3: Validation, Regularization & Callbacks

Estimated time: 6 hours

• Setting up validation datasets
• Applying regularization techniques to prevent overfitting
• Using callbacks such as EarlyStopping
• Training and tuning models on the Iris dataset

Module 4: Model Persistence & Advanced Structures

Estimated time: 6 hours

• Saving and loading models using Keras
• Choosing between full model and weight-only saving
• Introduction to pretrained models
• Overview of advanced architectures: CNNs, RNNs, Transformers, and Autoencoders

Prerequisites

• Proficiency in Python programming
• Familiarity with basic machine learning concepts
• Understanding of fundamental linear algebra and calculus

What You'll Be Able to Do After

• Build and train neural networks using Keras and TensorFlow
• Apply convolutional neural networks to image classification tasks
• Use validation, regularization, and callbacks to improve model performance
• Save, load, and deploy trained models efficiently
• Understand advanced deep learning architectures and their use cases