Professional Course
Deep Learning Specialization
From perceptrons to transformers — the complete, Australian-built pathway into modern AI.
DEE

$99.00
- Full lifetime access
- Certificate of completion
- Access on all devices
- 35+ of content
- Downloadable resources
About This Course
Course Overview
Deep learning is the engine behind the technologies reshaping every industry — the vision systems scanning crops across the Wimmera, the fraud models protecting Australian banks, the language models transforming how Geelong businesses write, search and serve customers. Yet most courses either drown you in mathematics or hand you a black box you can use but never truly understand.
The Deep Learning Specialization is a complete, end-to-end pathway through modern neural networks: you begin with a single logistic-regression neuron built from scratch in NumPy, and finish with convolutional networks, sequence models and the attention mechanisms that power today's transformer architectures — deploying a working model of your own along the way. Every example, dataset, price and regulation is Australian: AUD cloud-compute budgeting, the Privacy Act 1988 and the Australian Privacy Principles, Australia's AI Ethics Principles, and case studies drawn from local agriculture, health and finance rather than US defaults you then have to translate.
Fully self-paced and delivered through our Moodle LMS as interactive lessons, each chapter combines short, focused reading with hands-on coding notebooks, knowledge checks, drag-and-drop exercises and branching real-world scenarios — so you're building, not just watching, from the very first session.
Learning Outcomes
What You Will Learn
Explain how neural networks actually learn — forward propagation, cost functions, gradient descent and backpropagation — and implement all of it from scratch in NumPy
Build, train and tune deep networks in TensorFlow/Keras and PyTorch, choosing sensibly between the two
Diagnose bias and variance, and fix underperforming models with regularisation, dropout, batch normalisation and data augmentation
Accelerate training with mini-batch gradient descent, momentum, RMSprop, Adam and learning-rate schedules
Run a machine-learning project like a professional — orthogonalised goals, single-number metrics, human-level baselines and disciplined error analysis
Design and train convolutional neural networks, and apply transfer learning to modest Australian-scale datasets
Implement object detection and face-recognition pipelines using YOLO-style detectors, siamese networks and triplet loss
Build recurrent networks, GRUs and LSTMs for sequence data, and use word embeddings for practical NLP tasks
Explain attention and the transformer architecture — the foundation of modern large language models — and fine-tune pre-trained models responsibly
Deploy a trained model to the cloud from the AWS Sydney region, compliant with the Privacy Act 1988 and Australia's AI Ethics Principles
Before You Start
Requirements
- Comfort with basic Python — variables, functions and loops
- Willingness to revisit high-school-level algebra — Week 1 includes a guided NumPy and mathematics refresher
- A computer with a modern browser and Python 3 — all core exercises run free in Google Colab, no GPU purchase required
- A reliable internet connection
- Optional for Weeks 4–5: roughly $10–$30 AUD for paid GPU hours — a free-tier pathway is fully supported
Ideal For
Who Is This Course For?
Developers levelling upCareer changersData professionalsTechnical founders & leads
At a Glance
Course Details
Skill Level
Intermediate → Advanced
Language
English
Certificate
Yes, on completion
Access
Life Time
Format
Online, self-paced
Duration
35+


