Bridge: Python & Data Structures for AI
Python for data work, NumPy, and the data structures and algorithms every later subject assumes.
The programme
Five semesters, from an optional bridge to a solo capstone. Every subject is listed here with its credits and hours.
The semester map
- Semester 0: Bridge (No credits · optional)
- Semester 1: Foundations (14 core credits)
- Semester 2: Deep learning core (12 core credits)
- Semester 3: Domain track (8 track credits)
- Language models track
- Computer vision track
- Elective: Probabilistic ML & Bayesian Methods (2 credits)
- Elective: Reinforcement Learning Foundations (2 credits)
- Semester 4: Domain track and capstone (16 credits)
- Capstone (8 credits)
Semester 0: Bridge
Optional refreshers in Python, mathematics and SQL before Semester 1. No credits.
Bridge: Maths Refresher
Algebra, calculus, vectors and probability at school level, so Semester 1's mathematics lands on firm ground.
Bridge: SQL Essentials
Tables, SELECT, joins and aggregation in SQLite, ready for Data Engineering & Databases.
Semester 1: Foundations
Machine learning from first principles, with the linear algebra, calculus, probability and statistics it rests on, classical AI, and the data engineering every ML system needs.
Linear Algebra & Calculus for ML
Vectors, matrices, eigenvalues and the SVD, then gradients, Jacobians and the chain rule, computed in NumPy.
Probability & Statistics for ML
Probability, distributions, estimation and hypothesis tests, each result derived and then checked by simulation.
Foundations of Machine Learning
How machines learn from data, from linear models to ensembles and clustering, each method derived and built from scratch.
Classical AI: Search, Logic & Reasoning
Search, logic, Bayesian networks and planning: the ideas AI was built on and modern agents still use.
Data Engineering & Databases for ML
The relational model, SQL, storage and indexing, and reproducible data pipelines for machine learning.
Semester 2: Deep learning core
Neural networks and deep learning, optimisation and training at scale, sequence models and Transformers, and responsible AI. Every lab trains real models in PyTorch on one CPU core.
Neural Networks & Deep Learning
Backpropagation by hand, your own autodiff engine, then PyTorch, convolutions and residual networks.
Optimisation & Training at Scale
Why training works and what it costs: SGD to Adam, schedules, mixed precision and parallelism.
Sequence Models & Transformers
From n-grams and recurrent networks to the Transformer, built piece by piece, with tokenisation and decoding.
Responsible AI, Ethics & Governance
Fairness, explainability, privacy and robustness as engineering, and how AI and personal data are governed.
Semester 3: Domain track
Choose one track, language models or computer vision, and take its two Semester 3 subjects. Two electives, probabilistic ML and reinforcement learning, broaden the core.
Language models track
NLP & Language Model Foundations
Tokenisers, n-gram and neural language models, and text tasks across many languages and scripts.
Building and Training LLMs from Scratch
A tokeniser, a pretraining corpus and a GPT-2-style model, trained and costed on one CPU core.
Computer vision track
Image Formation and Classical Computer Vision
Camera models, filtering, features, homographies, stereo and optical flow, built in NumPy and SciPy.
Deep Visual Recognition
CNNs and Vision Transformers, detection and segmentation, self-supervised learning and compression for a CPU.
Electives
Open to both tracks. 2 credits each.
Probabilistic ML & Bayesian Methods
Posteriors exact and approximate: Gaussian processes, variational inference, MCMC and calibrated uncertainty.
Reinforcement Learning Foundations
Bandits, temporal-difference learning, Q-learning, deep Q-networks and policy gradients, on environments you write.
Semester 4: Domain track and capstone
Finish your track with its two Semester 4 subjects and run a solo capstone project in it, in five milestones from proposal to final report and viva.
Language models track
Post-training, Alignment and Evaluation of LLMs
Fine-tuning, reward models, PPO, GRPO and DPO, then an honest evaluation of what changed.
LLM Applications: Retrieval, RAG and Agents
Search from first principles, retrieval-augmented generation that cites its sources, and agents that use tools safely.
Computer vision track
Deep Generative Models
Autoregressive models, VAEs, flows, GANs, diffusion and flow matching, each derived and built from scratch.
Vision-Language, 3D Vision and World Models
Image-text models, 3D reconstruction, radiance fields and world models, trained small on one CPU core.
Capstone
Capstone Project
A solo project in your track, in five graded milestones from proposal to report, demo and viva.
Two tracks
Language Models & LLM Engineering
From text processing and tokenisation to training a small GPT from scratch, then post-training, alignment and evaluation, and building retrieval-augmented and agentic applications.
- 1NLP & Language Model Foundations
- 2Building and Training LLMs from Scratch
- 3Post-training, Alignment and Evaluation of LLMs
- 4LLM Applications: Retrieval, RAG and Agents
Computer Vision & Generative Models
From image formation and classical vision to deep visual recognition, then generative models of images, and vision-language, 3D vision and world models.
- 1Image Formation and Classical Computer Vision
- 2Deep Visual Recognition
- 3Deep Generative Models
- 4Vision-Language, 3D Vision and World Models
How prerequisites work
- Required
- Finish this subject before you start. The subject builds directly on it.
- Corequisite
- Take this subject before or alongside, in the same semester. Chapters lean on it as you go.
- Recommended
- Helpful background. You can start without it if you are willing to fill the gaps.
Semester 0 · Bridge
- Bridge: Python & Data Structures for AINeeds nothing before it.
- Bridge: Maths RefresherNeeds nothing before it.
- Bridge: SQL EssentialsNeeds nothing before it.
Semester 1 · Foundations
- Linear Algebra & Calculus for MLNeeds nothing before it.
- Probability & Statistics for MLNeeds nothing before it.
- Foundations of Machine LearningNeeds nothing before it.
- Classical AI: Search, Logic & ReasoningNeeds nothing before it.
- Data Engineering & Databases for MLNeeds nothing before it.
Semester 2 · Deep learning core
- Neural Networks & Deep LearningNeeds Foundations of Machine Learning, Linear Algebra & Calculus for ML.
- Optimisation & Training at ScaleNeeds Linear Algebra & Calculus for ML; alongside Neural Networks & Deep Learning.
- Sequence Models & TransformersNeeds Foundations of Machine Learning, Linear Algebra & Calculus for ML; alongside Neural Networks & Deep Learning.
- Responsible AI, Ethics & GovernanceNeeds Foundations of Machine Learning; alongside Neural Networks & Deep Learning.
Semester 3 · Domain track
- NLP & Language Model FoundationsNeeds Sequence Models & Transformers, Neural Networks & Deep Learning.
- Building and Training LLMs from ScratchNeeds Sequence Models & Transformers, Optimisation & Training at Scale, Neural Networks & Deep Learning.
- Image Formation and Classical Computer VisionNeeds Linear Algebra & Calculus for ML, Neural Networks & Deep Learning.
- Deep Visual RecognitionNeeds Neural Networks & Deep Learning, Sequence Models & Transformers.
- Probabilistic ML & Bayesian MethodsNeeds Probability & Statistics for ML, Linear Algebra & Calculus for ML, Foundations of Machine Learning, Neural Networks & Deep Learning.
- Reinforcement Learning FoundationsNeeds Classical AI: Search, Logic & Reasoning, Probability & Statistics for ML, Neural Networks & Deep Learning.
Semester 4 · Domain track and capstone
- Post-training, Alignment and Evaluation of LLMsNeeds Building and Training LLMs from Scratch, Responsible AI, Ethics & Governance.
- LLM Applications: Retrieval, RAG and AgentsNeeds NLP & Language Model Foundations, Sequence Models & Transformers, Neural Networks & Deep Learning.
- Deep Generative ModelsNeeds Deep Visual Recognition, Probability & Statistics for ML.
- Vision-Language, 3D Vision and World ModelsNeeds Deep Visual Recognition, Sequence Models & Transformers.
- Capstone ProjectNeeds Foundations of Machine Learning, Linear Algebra & Calculus for ML, Probability & Statistics for ML, Classical AI: Search, Logic & Reasoning, Data Engineering & Databases for ML, Neural Networks & Deep Learning, Optimisation & Training at Scale, Sequence Models & Transformers, Responsible AI, Ethics & Governance.