Research & Academy

Learn AI the way engineers build it.

Start from the mathematics, understand each architecture, then ship real systems. The same teardowns of RAG pipelines, agent loops and LLM internals from our YouTube channel, organised into a path.

Learning tracks.

Four tracks that build on each other. Take them in order, or join where your knowledge ends.

Mathematics of AI

Foundation

The handful of ideas every model is built from, explained visually.

  1. Vectors, matrices and linear transformations
  2. Derivatives, gradients and the chain rule
  3. Probability, distributions and Bayes
  4. Optimisation and gradient descent

Neural networks & deep learning

Core

Build a network from scratch, then understand the architectures that followed.

  1. Perceptrons, activations and backpropagation
  2. CNNs for vision
  3. RNNs, LSTMs and sequence models
  4. Regularisation, normalisation and training at scale

LLMs & transformers

Advanced

How large language models read, attend, predict and get adapted to new tasks.

  1. Tokenization and embeddings
  2. Self-attention and the transformer block
  3. Pre-training, fine-tuning and LoRA
  4. Alignment, prompting and evaluation

Applied AI engineering

Professional

Turn models into products that are reliable, measurable and affordable.

  1. RAG pipelines and vector databases
  2. Agents, tools and function calling
  3. Evals, guardrails and monitoring
  4. Deployment, MLOps and cost control
For organisations

Corporate AI training.

Hands-on programmes for engineering teams, and clear AI literacy sessions for leaders. Delivered on-site or live online, built around your own data and use cases.

  • Executive AI briefings, half day
  • Generative AI for developers, 3 days
  • RAG and agents bootcamp, 5 days
  • Custom curriculum and assessments
Plan a programme
How every network learns: one step of gradient descent

θt+1 = θt − η ∇θ L(θt)

θ
The model's weights
η
Learning rate: how big a step to take
∇L
Gradient of the loss: which way is downhill

Start with the free lessons.

Our explainers and glossary are open to everyone. Training for teams is one message away.