LEARNING PATHS
Work through a topic, step by step
A learning path is an ordered selection of articles from this wiki. Stop whenever you like — your progress is saved.
OVERVIEW
Getting started: understanding AI
What AI is, what it does, where it goes wrong, and which rules apply before you begin.
about 2 hours
7 articles
- 01Introduction3 min
- 02What artificial intelligence is2 min
- 03What AI can do today2 min
- 04Terms and model types2 min
- 05Getting started safely2 min
- 06Prompting basics2 min
- 07Common pitfalls · optional2 min
OVERVIEW
Applying: at home and at work
From your first task of your own to choosing a tool, with templates to take away.
about 3 hours
7 articles
- 01Writing and drafting2 min
- 02Learning and research2 min
- 03Planning and organising2 min
- 04Prompt templates2 min
- 05Sales1 min
- 06Choosing tools2 min
- 07AI at home · optional2 min
IN PRACTICE
Accountability: law and evidence
Roles, duties, deadlines and the evidence an audit actually asks for.
about 3 hours
7 articles
- 01The EU AI Act in outline2 min
- 02Duties by role2 min
- 03Deadlines and transitional rules2 min
- 04Legal bases2 min
- 05Logging2 min
- 06Preparing for audit1 min
- 07Copyright and AI · optional2 min
OVERVIEW
Prompting, properly
From a first instruction through the patterns to diagnosing why a prompt fails.
about 2 hours
5 articles
- 01Prompting basics2 min
- 02Prompt templates2 min
- 03Prompt patterns2 min
- 04Why a prompt fails2 min
- 05Prompt library · optional1 min
IN PRACTICE
Mathematical foundations
Vectors, derivatives, probability and information: the four tools everything else rests on.
about 4 hours
8 articles
- 01Maths overview3 min
- 02Linear algebra3 min
- 03Derivatives2 min
- 04Probability2 min
- 05Statistics2 min
- 06Information theory2 min
- 07Optimisation2 min
- 08Numerics · optional2 min
IN PRACTICE
Machine learning from the ground up
From linear regression to honest evaluation: the core that has held for fifty years.
about 5 hours
12 articles
- 01ML overview2 min
- 02Supervised learning2 min
- 03Regression2 min
- 04Loss functions2 min
- 05Gradient descent2 min
- 06Regularisation2 min
- 07Bias and variance2 min
- 08Cross-validation2 min
- 09Metrics2 min
- 10Feature engineering2 min
- 11Trees and ensembles2 min
- 12Clustering · optional2 min
IN DEPTH
Understanding language models
How a matrix multiplication becomes a language model, computed layer by layer.
about 5 hours
12 articles
- 01Neural networks2 min
- 02Tokenisation2 min
- 03Embeddings2 min
- 04Attention2 min
- 05The transformer architecture2 min
- 06How a language model is trained2 min
- 07RLHF and alignment2 min
- 08Sampling and temperature2 min
- 09Context windows in depth2 min
- 10Why models hallucinate2 min
- 11Reasoning models · optional2 min
- 12Mixture of experts · optional2 min
IN PRACTICE
Computer vision with OpenCV
How an image becomes numbers, what OpenCV solves classically, and where neural networks are needed.
about 6 hours
12 articles
- 01Computer vision2 min
- 02Images as data2 min
- 03OpenCV2 min
- 04Preprocessing with OpenCV2 min
- 05Edges and contours2 min
- 06Keypoints2 min
- 07Convolution and CNNs2 min
- 08Object detection2 min
- 09Segmentation2 min
- 10OCR2 min
- 11Vision transformers · optional2 min
- 12Face recognition · optional2 min
IN DEPTH
Hardware and operations
Why graphics cards, what memory bandwidth costs, and how a model reaches production.
about 4 hours
10 articles
- 01Why GPUs2 min
- 02GPU architecture2 min
- 03Memory and bandwidth2 min
- 04Training versus inference2 min
- 05Quantisation2 min
- 06Speeding up inference2 min
- 07Serving models2 min
- 08MLOps2 min
- 09Energy and sustainability2 min
- 10Edge AI · optional2 min
IN PRACTICE
Data: the craft
Collecting, labelling, checking, and the mistakes that devalue every later number.
about 3 hours
7 articles
- 01Building datasets2 min
- 02Measuring data quality2 min
- 03Annotation and labelling2 min
- 04Leakage2 min
- 05Bias in data2 min
- 06Reading benchmarks2 min
- 07Synthetic data · optional2 min
IN PRACTICE
The EU AI Act, compact
Who the law affects, which class a task falls into, which duties follow, and from when.
about 3 hours
6 articles
- 01The EU AI Act in outline2 min
- 02Risk classes2 min
- 03Duties by role2 min
- 04Deadlines and transitional rules2 min
- 05AI literacy2 min
- 06Approval process · optional2 min
IN PRACTICE
Data protection for AI
Which basis holds, where the data sits, who processes it, and what gets logged.
about 3 hours
6 articles
- 01Data protection basics2 min
- 02Legal bases2 min
- 03Processing on behalf2 min
- 04Data residency2 min
- 05Logging2 min
- 06Leakage · optional2 min
IN PRACTICE
Introducing AI in a team
From comparing tools through the pilot to rollout, with the numbers a decision rests on.
about 4 hours
8 articles
- 01Choosing tools2 min
- 02Evaluating new tools1 min
- 03Piloting1 min
- 04Metrics1 min
- 05Rollout1 min
- 06Training and enablement1 min
- 07Approval process2 min
- 08Reporting problems · optional2 min
OVERVIEW
AI for leaders
What has to be decided, which numbers carry, and which duties sit with leadership.
about 2 hours
6 articles
- 01What AI can do today2 min
- 02The EU AI Act in outline2 min
- 03AI literacy2 min
- 04What AI costs2 min
- 05Metrics1 min
- 06Rollout · optional1 min
IN PRACTICE
Law firm: AI in client work
Confidentiality, the duty to check and liability first, then the tasks that actually hold up.
about 5 hours
8 articles
- 01Law firm3 min
- 02Legal and compliance1 min
- 03Getting started safely2 min
- 04Prompt templates2 min
- 05RAG in depth2 min
- 06Copyright and AI2 min
- 07Processing on behalf2 min
- 08Logging · optional2 min
IN PRACTICE
Speech and audio
Recognising, understanding, generating: what speech technology does today and where it fails.
about 3 hours
5 articles
- 01NLP basics2 min
- 02Speech recognition2 min
- 03Speech synthesis3 min
- 04Audio as a signal2 min
- 05Multimodal models · optional2 min
IN DEPTH
Making your own documents usable
Retrieval, embeddings, vector search, and the measurement that shows whether an answer is covered.
about 5 hours
7 articles
- 01Embeddings2 min
- 02Vector databases2 min
- 03RAG in depth2 min
- 04Context windows in depth2 min
- 05Metrics2 min
- 06Why models hallucinate2 min
- 07Logging · optional2 min