Specialisation
Online Deep Learning Course
Backpropagation, CNNs, transformers and fine-tuning — taught as engineering, with models that end up behind an API rather than in a notebook.
- weeks
- 10weeks
- frameworks
- 2frameworks
- deployed model
- 1deployed model
What you get
How this one is different
Build the gradient yourself
Week one is a network in NumPy with a hand-written backward pass. Everything after that is easier for it.
PyTorch and TensorFlow
You learn both idioms, because job listings do not agree on one and the concepts transfer either way.
Transformers in depth
Attention, positional encoding, tokenisation, fine-tuning a small model and evaluating generated output.
Ships to production
The final project is served behind an API with monitoring — a model nobody can call is not finished.
Detail
What you will build
- 01A neural network from scratch in NumPy
- 02An image classifier with a CNN
- 03A sequence model for time-ordered data
- 04A fine-tuned transformer for a domain task
- 05A retrieval-augmented question answering system
- 06A deployed inference endpoint with logging
Questions
Before you start
Comfort with linear algebra and derivatives. We revise both in context during week one rather than assuming them.
JOIN NOW
Begin the first module
It is free, it is the real curriculum, and if it is not for you, you have lost nothing but an evening.
Join any time · Build AI skills at your pace