Glossary

AI (Artificial Intelligence)

Technology that lets computers perform tasks that normally need human thinking, like understanding language or spotting patterns in data.

GenAI (Generative AI)

AI that creates brand-new content — such as text, images, code, or summaries — based on the instructions you give it. ChatGPT and Claude are examples.

LLM (Large Language Model)

An AI model trained on huge amounts of text so it can understand and generate human-like language. Claude, GPT, and Llama are all LLMs.

RAG (Retrieval-Augmented Generation)

A technique where an AI first looks up relevant information from your own documents or database, then uses that information to write an accurate, grounded answer — instead of relying only on what it memorized during training.

Prompt Engineering

The practice of carefully wording instructions given to an AI model so it produces the most accurate and useful response.

Embedding

A way of converting text (or images) into a list of numbers that captures its meaning, so a computer can measure how similar two pieces of content are.

Vector Search / Vector Database

A search method that finds content by meaning rather than exact keywords, by comparing embeddings. A vector database is where those embeddings are stored for fast lookup.

LangChain / LangGraph

Developer frameworks for building AI applications. LangChain provides building blocks for connecting an LLM to data and tools; LangGraph adds the ability to design multi-step, decision-making workflows (an "agent") on top of that.

AI Agent

An AI system that can plan and take a sequence of actions on its own — such as deciding which data source to query or which tool to call — to complete a task, rather than just answering a single question.

Text-to-SQL

AI that converts an everyday question (e.g. "how many customers churned last month?") into a working SQL database query automatically.

SQL

A standard language used to ask questions of and retrieve data from a database.

Python / R

Popular programming languages used for data analysis, automation, and statistics. Python is more general-purpose; R is more specialized for statistics.

Pandas

A widely used Python library for cleaning, organizing, and analyzing tabular data (like a spreadsheet, but programmable).

API (Application Programming Interface)

A defined way for two pieces of software to talk to each other — for example, letting a website request data from a server.

FastAPI

A Python framework for quickly building APIs, often used as the backend that connects an AI model to a website.

BigQuery

Google's cloud-based data warehouse — a place to store and quickly analyze very large datasets (millions or billions of rows) using SQL.

DuckDB / Supabase

Modern lightweight database tools. DuckDB runs fast analytical queries locally without needing a server; Supabase is a hosted database platform built on PostgreSQL.

Pinecone

A managed vector database service, commonly used to power semantic (meaning-based) search in AI applications like RAG.

Airflow

A tool for scheduling and automating multi-step data workflows (e.g. "every night, pull new data, clean it, then update the dashboard") so they run reliably without manual effort.

Docker

A tool that packages an application with everything it needs to run, so it behaves the same way on any computer or server.

VPS (Virtual Private Server)

A rented virtual computer that runs continuously in the cloud, often used to host applications or automated jobs 24/7.

Data Pipeline

An automated sequence of steps that moves data from its raw source, cleans and transforms it, and delivers it somewhere useful, like a dashboard or database.

Power BI / Tableau / Looker

Business intelligence tools used to turn raw data into interactive charts and dashboards that non-technical people can explore.

Streamlit

A Python tool for quickly building interactive web apps and dashboards for data, without needing traditional web development skills.

Dashboard

A visual, often interactive, summary screen that shows key data and metrics at a glance.

Descriptive / Diagnostic / Predictive Analytics

Three stages of data analysis: descriptive explains "what happened," diagnostic explains "why it happened," and predictive estimates "what's likely to happen next."

Churn (Customer Churn)

The rate at which customers stop using a company's product or service. Churn analysis tries to identify who is likely to leave and why.

KPI (Key Performance Indicator)

A specific, measurable number used to track how well something is performing against a goal (e.g. monthly sales growth).

CEFR

Common European Framework of Reference for Languages — an international standard for rating language proficiency, from A1 (beginner) to C2 (mastery).

TPKA (Tes Potensi Kognitif Akademik)

An Indonesian academic aptitude test that measures verbal, quantitative, and logical reasoning ability, often required for graduate school admission.

GPA (Grade Point Average)

A standard measure of academic performance, typically on a 4.00 scale, averaged across all courses taken.