Certification programme
Everything on the modelling side, in one programme. Data science end to end, the machine learning algorithms understood rather than imported, and applied AI through to generative models and guardrails.
What you will be able to do
Everything included
27 modules · 140 topics · 557 concepts
This is our Data Science course in full — the same 10 modules and 210 concepts, not an abridged version.
Python for data work
NumPy
pandas
Reproducibility
Version control for data work
You can manipulate real datasets fluently in pandas and NumPy.
SQL for data science
APIs and scraping
File formats at scale
Data warehouses and lakes
Pipelines
You can assemble a dataset from databases, APIs and files yourself.
Descriptive and exploratory
Probability
Estimation
Hypothesis testing
Experiment design
Causal thinking
You can quantify uncertainty and design an experiment that answers a question.
The learning framing
Preprocessing
Regression and classification
Evaluation
Ensembles
Interpretability
You can build and validate a model without leaking information into it.
Clustering
Dimensionality reduction
Anomaly detection
Recommenders
You can segment a population and build a working recommender.
Neural network basics
Training
Regularisation
Convolutional networks
Sequence models
Practicalities
You can train a network and diagnose why one is not learning.
Classical NLP
Embeddings
Transformers in practice
Large language models
Text tasks
You can solve a text problem with embeddings or a fine-tuned model.
Why single-machine fails
Distributed processing
Cloud data platforms
Streaming in outline
You can process data that does not fit on one machine.
Packaging
Serving
MLOps
Monitoring
Cost and reliability
You can deploy, monitor and retrain a model responsibly.
Responsible use
Communicating results
Working with stakeholders
Portfolio
Interview preparation
You can present a result to non-technical stakeholders and defend your choices.
This is our Machine Learning course in full — the same 9 modules and 184 concepts, not an abridged version.
Types of learning
The core vocabulary
The fundamental trade-off
Evaluation discipline
When not to use ML
You can state any problem as a learning task and identify what could go wrong.
Linear algebra
Calculus for optimisation
Gradient descent
Probability
Statistics for ML
You can read an algorithm description and follow what it is optimising.
Exploratory analysis
Missing data
Encoding
Scaling
Feature engineering
Pipelines
You can build a preprocessing pipeline that does not leak information.
Linear regression
Assumptions
Polynomial and non-linear
Regularisation
Regression metrics
You can fit, regularise and interpret a regression model.
Logistic regression
K-nearest neighbours
Naive Bayes
Support vector machines
Decision trees
Classification metrics
You can build a classifier and choose a threshold defensibly.
Why ensembles work
Bagging and random forests
Boosting
Stacking
Interpretability
You can train and tune a gradient-boosted model and explain its predictions.
Clustering
Evaluating clusters
Dimensionality reduction
Anomaly detection
Association rules
You can cluster and reduce dimensionality, and judge whether the result is meaningful.
Cross-validation
Hyperparameter search
Imbalanced data
Model selection
Common failures
You can tune a model honestly and report a number you would defend.
Persisting a model
Serving
Monitoring
Experiment tracking
Ethics and risk
You can serve a model behind an API and monitor it for drift.
This is our Artificial Intelligence course in full — the same 8 modules and 163 concepts, not an abridged version.
Definitions
A short history
Capabilities and limits
The ecosystem
Setting up
You can place any AI system in context and say what it can and cannot do.
The neuron
Activations
Forward propagation
Loss functions
Backpropagation
Training loop
You can implement forward and backward passes and explain every step.
Optimisation in practice
Regularisation
Initialisation and stability
Data handling
Debugging training
You can diagnose a failing training run and fix it.
Images as data
Convolutional networks
Architectures
Transfer learning
Beyond classification
Evaluation
You can build an image model using transfer learning and evaluate it properly.
Text preprocessing
Classical approaches
Embeddings
Transformers
Using pretrained models
You can solve a language task with the right tool for its size and budget.
How LLMs work
Prompting as engineering
Retrieval augmented generation
Generative models beyond text
Evaluating generative output
You can build a grounded LLM application and evaluate its output.
Serving models
Working with model APIs
System design
Monitoring
Edge and on-device
You can ship an AI feature with sensible cost, latency and failure behaviour.
Bias and fairness
Privacy
Security of AI systems
Transparency
Keeping current
You can assess an AI system for risk and explain your assessment.
2 builds you keep
Forecasting, churn, credit scoring with explanations, a recommender, a defect detector, a document pipeline, a grounded assistant and a fairness audit, among others.
A system that uses classical modelling where classical modelling wins and a neural approach where it does not, with the comparison documented rather than assumed.
Hands on with
Before you enrol
There is deliberate overlap and we would rather be straight about it. The Data Science track covers machine learning at working depth; the Machine Learning track then derives the same algorithms properly and goes into ensembles, tuning and validation discipline. If you want one and not the other, take them separately.
Identical. All four pages are generated from one source, so a module here is the same module there. Nothing is trimmed to fit the package.
Yes, and for most people it is the better one. Machine Learning alone covers the models most businesses actually need. This programme is for people who want the whole modelling side including vision, language and generative AI.
Comfortable Python and some SQL. If you do not have that, Data Science with Python starts a step earlier and gets you here.
Not one you own. The AI track uses free cloud notebook GPUs and transfer learning, which keeps training runs short enough for free tiers.
Where this leads
One skill on a CV is a hobby. These are the courses students take alongside this one, in the order that builds an actual role.
The analyst toolkit and the scientist toolkit, in the order employers actually hire for. Get employable as an analyst first, then add the modelling and deployment that moves you up.
Statistics, machine learning and big data tooling, taught through projects that end in deployed models rather than notebooks. Six months, with placement preparation built in.
The full route for someone who has never programmed: Python properly first, then the whole data science programme built on it. Same syllabus as both courses taken separately, in the order that actually works.