Data & AI1:1 mentoring

Deep Learning

Understand neural networks from the neuron up, then build CNNs, sequence models and transformers in PyTorch.

Level:: Intermediate to advanced
Format:: 1:1 online mentoring

Free, no-obligation chat · Affordable plans that fit your budget · Start this week

1:1 personal attentionOne mentor, one student, at your pace.
Flexible timingsSessions that fit your week and time zone.
Affordable plansPlans that fit your budget.
Practise between sessionsShort problems that build confidence.

About this course

Learn Deep Learning online

Deep learning powers image recognition, speech, translation and modern language models. It can feel like magic until you see the pieces: a neuron is a weighted sum, training is repeated gradient steps, and backpropagation is the chain rule. Beginners often struggle with tensor shapes, exploding losses, overfitting and knowing what to change when a network does not learn.

This course follows the standard deep learning syllabus: neural network basics, activation and loss functions, backpropagation, optimisers, regularisation, convolutional networks, recurrent networks and LSTMs, attention and transformers, transfer learning, generative models at an introductory level, and practical training in PyTorch. Each module lists its topics and what you will be able to do.

Learning is through affordable 1:1 mentoring, where a mentor explains the maths and the code together and reviews your training runs with you. Message us on WhatsApp or email, or fill in the short form, for a free, no-obligation chat. We will suggest a plan that fits your budget and flexible timings, and you can start this week.

Who this course is for

Students taking a deep learning or neural networks course
Machine learning learners ready for neural networks
Software engineers who want to build models with PyTorch
Researchers and analysts working with images, text or sequences
Learners who want to understand how modern language models work

Outcomes

What you'll learn

Explain how a neural network computes a prediction and learns from its error
Derive and implement backpropagation for a small network
Choose activation functions, loss functions and optimisers sensibly
Build and train models in PyTorch using tensors, autograd, datasets and data loaders
Design convolutional networks for image classification
Use RNNs and LSTMs for sequence data
Explain attention and the transformer architecture
Apply transfer learning and regularisation to get good results with limited data

Skills you'll gain

PyTorch
Neural networks
Backpropagation
Optimisation
CNNs
RNNs and LSTMs
Transformers
Transfer learning
Regularisation
Tensors
Training and debugging
Python

Detailed syllabus

Deep Learning syllabus

Every module lists its topics and what you will be able to do. Your mentor plans sessions around them and your own goals.

Module 1
Foundations and the maths you need
The linear algebra, calculus and Python tools that neural networks rest on.
- What deep learning is and how it differs from classical machine learning
- Vectors, matrices and tensors
- Matrix multiplication
- Derivatives, partial derivatives and the chain rule
- Gradients and the idea of descent
- NumPy review
- Setting up PyTorch
By the end of this module you can
- Compute gradients of simple functions by hand
- Use tensors in NumPy and PyTorch
Covered in 1:1 mentoring sessions
Module 2
Neural network basics
From a single neuron to a multi-layer network.
- The perceptron
- Weights, bias and the weighted sum
- Activation functions: sigmoid, tanh, ReLU
- Layers and the forward pass
- Multi-layer perceptrons
- The universal approximation idea
- Network capacity
- Loss functions: MSE and cross-entropy
By the end of this module you can
- Compute a forward pass through a small network by hand
- Choose an activation and a loss function for a task
Covered in 1:1 mentoring sessions
Module 3
Training: backpropagation and optimisers
How a network learns from data.
- Gradient descent, stochastic and mini-batch
- Backpropagation with the chain rule
- Computational graphs
- Learning rate and its effect
- Momentum, RMSProp and Adam
- Weight initialisation
- Vanishing and exploding gradients
- Learning rate schedules
By the end of this module you can
- Implement backpropagation for a small network
- Pick an optimiser and a learning rate and diagnose a loss that will not fall
Covered in 1:1 mentoring sessions
Module 4
PyTorch essentials
The practical tools for building and training models.
- Tensors, devices and dtypes
- Autograd and requires_grad
- nn.Module and nn.Sequential
- Dataset and DataLoader
- The training loop
- Evaluation mode and torch.no_grad
- Saving and loading models
- Using a GPU when available
By the end of this module you can
- Write a complete PyTorch training and evaluation loop
- Debug tensor shape errors quickly
Covered in 1:1 mentoring sessions
Module 5
Regularisation and good training practice
Making networks generalise instead of memorise.
- Overfitting and underfitting
- Train, validation and test sets
- L2 weight decay
- Dropout
- Batch normalisation
- Data augmentation
- Early stopping
- Hyperparameter search
- Reproducibility with seeds
By the end of this module you can
- Read training and validation curves to find overfitting
- Apply dropout, weight decay and augmentation to improve generalisation
Covered in 1:1 mentoring sessions
Module 6
Convolutional neural networks
Networks designed for images.
- Why fully connected layers struggle with images
- Convolution, filters and feature maps
- Stride and padding
- Pooling
- Building a CNN in PyTorch
- Classic architectures: LeNet, VGG and ResNet
- Residual connections
- Visualising what filters learn
By the end of this module you can
- Compute output sizes for convolution and pooling layers
- Train a CNN to classify images
Covered in 1:1 mentoring sessions
Module 7
Transfer learning and computer vision tasks
Reusing pretrained networks and meeting other vision problems.
- Pretrained models and feature extraction
- Fine-tuning
- Freezing layers
- Image normalisation and preprocessing
- Object detection overview
- Segmentation overview
- Evaluating vision models
By the end of this module you can
- Fine-tune a pretrained network on a small dataset
- Explain the difference between classification, detection and segmentation
Covered in 1:1 mentoring sessions
Module 8
Sequence models: RNNs and LSTMs
Networks that remember what came before.
- Sequence data: text, audio and time series
- Recurrent neural networks
- Backpropagation through time
- Vanishing gradients in sequences
- LSTM and GRU cells
- Embeddings
- Bidirectional models
- Sequence-to-sequence models
By the end of this module you can
- Train an LSTM on a text or time series task
- Explain how gates help an LSTM keep information
Covered in 1:1 mentoring sessions
Module 9
Attention and transformers
The architecture behind modern language models.
- Limits of recurrent models
- Attention: queries, keys and values
- Scaled dot-product attention
- Multi-head attention
- Positional encoding
- The transformer encoder and decoder
- Training transformers
- Why transformers scale
By the end of this module you can
- Compute self-attention for a short sequence by hand
- Describe each block of a transformer and its role
Covered in 1:1 mentoring sessions
Module 10
Pretrained language models and fine-tuning
Using and adapting large pretrained models.
- Tokenisation
- Pretraining and fine-tuning
- Encoder, decoder and encoder-decoder models
- Using pretrained models for classification
- Parameter-efficient fine-tuning idea
- Evaluation of language models
- Limits and risks
By the end of this module you can
- Fine-tune a small pretrained model for a text task
- Explain pretraining versus fine-tuning
Covered in 1:1 mentoring sessions
Module 11
Generative models and beyond
An introduction to models that create new data.
- Autoencoders
- Variational autoencoders overview
- Generative adversarial networks overview
- Diffusion models overview
- Reinforcement learning at a glance
- Responsible use of generative models
By the end of this module you can
- Build a simple autoencoder
- Describe in plain words how GANs and diffusion models generate data
Covered in 1:1 mentoring sessions
Module 12
Capstone project and deployment
Taking a deep learning project from idea to a working demo.
- Choosing a problem and data
- Setting a baseline
- Experiment tracking
- Error analysis
- Exporting a trained model
- Serving a model behind a simple API
- Writing up results
By the end of this module you can
- Deliver a complete deep learning project with a clear baseline
- Explain your results and their limits
Covered in 1:1 mentoring sessions

Ways to learn

How you can learn Deep Learning

Deep Learning mentor, 1:1 online

A mentor who knows this course plans sessions around your goal, explains each idea with fresh examples and checks your understanding. Personal attention, flexible timings in your time zone and affordable plans that fit your budget.

Talk to our team

How it works

Your learning journey

Four simple steps to start Deep Learning with a mentor. A real person replies to you; there is no booking system.

Step 1: Talk to us
Message us on WhatsApp or email, or fill in the short form. The first chat is free, with no obligation.
Step 2: Get your plan
We listen to your goal, level and deadlines, then suggest a plan that fits your budget and timings.
Step 3: Learn with your mentor
Online 1:1 sessions with a mentor who knows your course, explains clearly and checks your understanding.
Step 4: Practise & track progress
Keep going between sessions with practice, and watch your confidence grow week by week.

Learn by doing

Projects and practice

Handwritten digit recogniser
Build a neural network from scratch in NumPy, then rebuild it in PyTorch and compare the two.
Image classifier with transfer learning
Fine-tune a pretrained CNN on a small image dataset and analyse its mistakes.
Text sentiment with an LSTM
Train a recurrent model on review text and compare it with a simple baseline.
Mini attention block
Implement self-attention from scratch and use it in a small sequence model.
Capstone demo
Choose your own problem, train and evaluate a model, and serve it behind a simple web API.

Free practice on MentorNxt

Your progress

What you'll be able to do

Milestones from the start, middle and end of the syllabus.

Compute gradients of simple functions by hand
Early on
Fine-tune a pretrained network on a small dataset
Halfway
Deliver a complete deep learning project with a clear baseline
By the end

FAQs

Deep Learning: frequently asked questions

Is there a self-study option?

Not for deep learning itself. Our free self-study lessons exist only for Java and Python programming, so this course is taught through 1:1 mentoring. You can use our free Python lessons to prepare.

What should I know before starting?

Comfortable Python and NumPy, plus basic machine learning ideas such as training and test data, and overfitting. Our machine learning course is the natural step before this one. Basic calculus and linear algebra help, and your mentor will revisit them as needed.

Do I need a GPU?

Not to start. Small networks train on an ordinary laptop, and your mentor can show you free hosted notebooks with a GPU for larger exercises.

PyTorch or another framework?

We teach PyTorch, which is widely used in research and industry and makes the ideas easy to see in code. The concepts transfer to any framework.

Will I learn how large language models work?

Yes. You will study attention and the transformer architecture, then see how pretrained models are fine-tuned. Our artificial intelligence and generative AI course covers building applications on top of such models.

How much does mentoring cost?

1:1 mentoring is personal online tutoring and is affordable, with plans that fit your budget. Tell us your goal and how often you would like sessions, and our team will suggest an option that works for you in a free, no-obligation chat.

How do I get started?

Message us on WhatsApp or email, or fill in the short form on this page. Our team replies, talks through your goal and level, and arranges a mentor at flexible timings in your time zone. You can start this week.

Free, no-obligation chat

Talk to our team about Deep Learning

Tell us what you need. Our support team replies within one working day, by email or WhatsApp, whichever you prefer, and arranges your mentor.

Prefer to talk first?

Message the support team directly. We usually reply the same day.

Chat on WhatsApp Email us

Our mentors teach and explain. They never complete graded work, tests or assessments for students.

Keep going

Syllabus references

Last reviewed9 November 2026

Ready to start Deep Learning?

Message us on WhatsApp or email, or fill in the short form. Our team replies with a plan that fits your goal and budget, and you can start this week.

Fill in the short form

WhatsApp Email us

Talk to us about a mentor

WhatsApp Email us

Learn Deep Learning online

Who this course is for

What you'll learn

Skills you'll gain

Deep Learning syllabus

Foundations and the maths you need

Neural network basics

Training: backpropagation and optimisers

PyTorch essentials

Regularisation and good training practice

Convolutional neural networks

Transfer learning and computer vision tasks

Sequence models: RNNs and LSTMs

Attention and transformers

Pretrained language models and fine-tuning

Generative models and beyond

Capstone project and deployment

How you can learn Deep Learning

Deep Learning mentor, 1:1 online

Your learning journey

Step 1: Talk to us

Step 2: Get your plan

Step 3: Learn with your mentor

Step 4: Practise & track progress

Projects and practice

Handwritten digit recogniser

Image classifier with transfer learning

Text sentiment with an LSTM

Mini attention block

Capstone demo

Free practice on MentorNxt

What you'll be able to do

Deep Learning: frequently asked questions

Talk to our team about Deep Learning

Related courses

Machine Learning with Python

Artificial Intelligence and Generative AI

Data Science with Python

Python Programming

Discrete Mathematics

Syllabus references

Ready to start Deep Learning?

Learn Deep Learning online

Who this course is for

What you'll learn

Skills you'll gain

Deep Learning syllabus

Foundations and the maths you need

Neural network basics

Training: backpropagation and optimisers

PyTorch essentials

Regularisation and good training practice

Convolutional neural networks

Transfer learning and computer vision tasks

Sequence models: RNNs and LSTMs

Attention and transformers

Pretrained language models and fine-tuning

Generative models and beyond

Capstone project and deployment

How you can learn Deep Learning

Deep Learning mentor, 1:1 online

Your learning journey

Step 1: Talk to us

Step 2: Get your plan

Step 3: Learn with your mentor

Step 4: Practise & track progress

Projects and practice

Handwritten digit recogniser

Image classifier with transfer learning

Text sentiment with an LSTM

Mini attention block

Capstone demo

Free practice on MentorNxt

What you'll be able to do

Deep Learning: frequently asked questions

Talk to our team about Deep Learning

Related courses

Machine Learning with Python

Artificial Intelligence and Generative AI

Data Science with Python