Data and AI terms, in plain words.

AI ActDutch: AI-verordening (AI Act)

The AI Act is the European regulation on artificial intelligence, Regulation (EU) 2024/1689.

AI agentDutch: AI-agent

An AI agent is a system in which a language model decides which steps to take and which tools to call to reach a goal, and repeats until it is done or stops.

AI harnessDutch: AI-harness

An AI harness is the code and configuration around a model that makes it usable in production: prompts, tools, permissions, memory, logging and automatic checks.

Asset BundlesDutch: Asset Bundles

Databricks Asset Bundles describe jobs, pipelines and other Databricks resources as code in a configuration file.

Data residencyDutch: Dataresidentie

Data residency is the requirement or choice that data is stored and processed in a particular country or region.

Delta LakeDutch: Delta Lake

Delta Lake is an open table format that adds a transaction log to files in cloud storage.

Document AIDutch: Document-AI

Document AI reads documents such as invoices, forms and contracts and turns them into structured data.

DTAPDutch: DTAP-straat

DTAP stands for development, test, acceptance and production: the environments a change passes through before it reaches users.

Evaluation setDutch: Evaluatieset

An evaluation set is a collection of real examples with the answers or behaviour you expect, used to measure a model or an AI system.

Fine-tuningDutch: Finetuning

Fine-tuning continues the training of an existing model on a smaller set of your own examples.

GuardrailsDutch: Guardrails

Guardrails are checks around an AI system that limit what goes in and what comes out.

InferenceDutch: Inferentie (inference)

Inference is using a trained model to produce an answer, as opposed to training it.

LakehouseDutch: Lakehouse

A lakehouse keeps data in open formats on low-cost cloud storage and adds the management features of a data warehouse on top, such as tables, transactions and access control.

MCP (Model Context Protocol)Dutch: MCP (Model Context Protocol)

The Model Context Protocol is an open standard for connecting AI applications to tools and data sources.

Open-weight modelDutch: Open-weightmodel

An open-weight model is a model whose trained parameters (the weights) are published, so you can download and run it yourself.

RAG (retrieval-augmented generation)Dutch: RAG (retrieval-augmented generation)

RAG first searches a body of your own documents for passages that match a question, then gives those passages to a language model to write the answer.

Self-hosted LLMDutch: Zelf gehost taalmodel

A self-hosted large language model runs on infrastructure you or your partner control, such as your own servers or a cloud region you choose, instead of behind a vendor's API.

Unity CatalogDutch: Unity Catalog

Unity Catalog is the governance layer of the Databricks platform.

Vector searchDutch: Vectorzoeken

Vector search finds items by meaning instead of exact words.

Voice agentDutch: Voice agent

A voice agent is an AI system that holds a spoken conversation, usually over the phone.