Glossary

Machine learning is the branch of artificial intelligence in which a system gets better at a task by finding patterns in data, instead of following rules a programmer wrote out by hand.

A neural network is a model made of layers of simple connected units whose connection strengths are tuned, through training, until the whole network maps inputs to the outputs you want.

Deep learning is machine learning with many-layered neural networks that learn their own features directly from raw data, rather than relying on features a human picks out first.

Generative AI is a class of models that produce new content — text, images, audio or code — by learning the patterns of a huge body of examples and then sampling fresh combinations from them.

A large language model is a neural network trained on vast amounts of text to predict the next piece of text, which turns out to be enough to make it write, summarise, translate and reason in language.

The Transformer is a neural-network architecture that processes a whole sequence at once and uses an 'attention' mechanism to weigh how much each part should influence every other part.

Prompt engineering is the practice of writing the input to a generative model carefully — instructions, context and examples — so that it reliably produces the output you actually want.

Retrieval-augmented generation is a pattern that fetches relevant documents at question time and puts them into a model's prompt, so its answer is grounded in specific, current sources rather than memory alone.

Natural language processing is the field of AI concerned with getting computers to work with human language — understanding it, generating it, translating it and pulling meaning from it.

Computer vision is the field of AI that gets machines to interpret images and video — detecting objects, classifying scenes, reading text in pictures and tracking what moves.

Agentic AI is the step from models that answer to systems that act: software built around a language model that plans a task, calls tools and APIs, checks its own progress and keeps going across multiple steps with limited human intervention.

MLOps is the engineering discipline of getting machine-learning models out of notebooks and into production — versioning, deploying, monitoring and retraining them — so they keep working after the world stops looking like the training data.

Responsible AI is the practice of building and deploying AI systems whose outcomes you can defend — fair, explainable, private and accountable — treating 'should we, and under which safeguards?' as an engineering requirement rather than an afterthought.

Data analysis is the process of examining data to answer a question — cleaning it, exploring it, and drawing conclusions you can act on and defend.

Data wrangling is the work of turning raw, messy data into a clean, consistent shape that analysis or a model can actually use.

Data visualization is the practice of representing data graphically so that patterns, comparisons and outliers become obvious to a human eye.

Data engineering is the discipline of building the pipelines and storage that move data reliably from where it is created to where it can be analysed or fed to a model.

Python is a general-purpose programming language prized for readable syntax and a vast ecosystem of libraries, which has made it the default language of data and AI.

Digital marketing is the practice of reaching and converting customers through digital channels — search, social, email and the web — measured and optimised with data.

SEO is the craft of being the answer: shaping a site's content, structure and technical health so that when someone searches for what you offer, search systems find you, understand you and rank you.

Marketing automation is software doing the follow-up a human team would forget: triggered, personalised sequences of emails, messages and audience updates that respond to what each contact actually does, at a scale no team could handle by hand.

Shared hosting places many customers' websites on one server that pools its processor, memory and disk between them, which makes it the cheapest way to put a site online — and the reason a busy neighbor can slow yours down.

A VPS is a software-defined slice of a physical server that behaves like a machine of its own — reserved CPU, memory and disk plus full administrative access — so you get root without paying for the whole box.

A dedicated server is an entire physical machine rented by one customer — no hypervisor, no co-tenants — chosen when raw performance, hardware-level control or strict isolation matters more than the flexibility of virtual machines.

Managed hosting is any hosting arrangement where the provider operates the server for you — updates, security, monitoring, backups and tuning — so what you are really buying is ops expertise attached to infrastructure, not the infrastructure itself.

Cloud hosting provisions computing resources — virtual machines, storage, databases — on demand through an API from a provider's pooled infrastructure, billed by usage rather than by rented box.

A web server is the program that listens for HTTP requests and answers them — serving files directly or passing the request to application code — with nginx and Apache as the long-standing workhorses of the role.

IaaS is the rawest cloud offering: virtual machines, storage volumes and networks rented as building blocks through an API, with everything above the infrastructure — operating system, runtime, application — assembled and operated by you.

PaaS is a cloud layer where you hand over code and the platform does the rest — provisioning, runtimes, scaling and deployment — trading infrastructure control for development speed.

Serverless computing runs code on demand without any server you rent or manage: functions execute when an event triggers them, scale to zero between requests, and bill by actual execution rather than reserved capacity.

A container packages an application together with its runtime environment — libraries, dependencies, configuration — into one portable unit that runs identically wherever a container engine exists, isolated by the host kernel rather than by a virtual machine.

Object storage keeps files as self-contained objects in flat buckets addressed over HTTP — no file system, no attached disk — scaling to practically unlimited size and billed by what you store and transfer.

Load balancing distributes incoming requests across multiple servers so that no single machine carries the traffic alone — the mechanism that turns 'add more servers' into an actual scaling strategy and single-server failure into a non-event.

A CDN is a network of servers spread across the world that keeps copies of your site's content close to visitors, so requests are answered from a nearby edge location instead of traveling to your origin server.

DNS is the internet's address book: the distributed system that translates human-readable domain names into the machine addresses and service pointers that actually route traffic — and the control panel where a domain's fate is decided.

TLS (still widely called SSL) is the protocol that encrypts traffic between browser and server and proves, via certificates, that the server really speaks for its domain — the mechanism behind the padlock and the S in HTTPS.

A domain name is a rented address in the internet's namespace — registered through a registrar for annual fees, controlled entirely by whoever holds the registrar account, and independent of wherever the website behind it happens to be hosted.

Caching stores the result of expensive work — a rendered page, a database query, a downloaded file — and serves the stored copy while it is still valid, trading a little freshness for most of a system's speed.

Uptime measures the share of time a service is reachable, quoted in 'nines'; an SLA (service level agreement) is the contract clause that defines how that share is counted and what compensation — usually service credit — a miss triggers.