NEW RELEASE: AI Training

Master the complete AI pipeline, from core theory to hardware deployment. Led by an embedded AI expert, in this course you will learn to build, train, and evaluate your own models on resource-constrained devices for real-world application.

Due to the length and depth of the course the delivery is split over a number of sessions and not delivered in a continuous block.

Module 1: Machine Learning Basics

• Supervised, unsupervised, reinforcement learning
• Regression and classification
• Loss functions, overfitting, regularization
• Model evaluation (cross-validation, metrics)

Module 2: Core Algorithms

• Decision Trees
• k-Nearest Neighbors
• Naive Bayes
• Support Vector Machines
• k-Means clustering

Module 3: Neural Networks & Deep Learning

• Perceptron and multi-layer networks
• Activation functions
• Backpropagation
• Intro to CNNs, RNNs

Module 4: Building & Using Pretrained Models

• Model construction with TensorFlow (Sequential API for CNNs)
• Building a simple CNN for image classification
• Loading and using a pretrained model (e.g. ResNet50 or MobileNet)
• Modifying the output layer for a custom dataset
• Training both models on the same dataset
• Comparing performance: accuracy, loss, and training time

Module 5: Deployment & Hardware Evaluation

• Exporting TensorFlow models (SavedModel, TFLite)
• Deploying models on Raspberry Pi (setup, dependencies, accelerators)
• Running inference with a pretrained ResNet and a custom CNN
• Measuring latency, memory usage, and power consumption
• Comparing performance: pretrained vs. user-built model on edge hardware