Glossary

AI systems that pursue a goal over multiple steps—planning, calling tools, and acting on results—rather than returning a single response to a single prompt.

AI Governance

AI governance encompasses the principles, policies, and practices that guide the responsible development and deployment of artificial intelligence systems.

AI Operating System

Learn what an AI Operating System (AI OS) is, how it works, and why modern teams are adopting it to to scale AI faster.

Compound AI System

A compound AI system combines multiple models, retrievers, and external tools to solve complex tasks better than monolithic models alone.

Context Window

The maximum amount of text, measured in tokens, that a language model can consider at once—covering the prompt, retrieved context, and the response it generates.

Data Lineage

Data lineage creates a comprehensive map of data's origin, movement, and transformations to ensure governance, accuracy, and compliance.

Data Loss Prevention (DLP)

Data Loss Prevention (DLP) stops sensitive data from leaving your environment. For AI workflows, DLP is the control that keeps customer PII, PHI, and financial records from being pasted into a public model.

Data Sovereignty

Data sovereignty is the legal requirement that digital data remains subject to the laws of the jurisdiction where it is collected.

Drift Monitoring

Drift monitoring is a crucial process in machine learning that tracks changes in data distributions over time, ensuring model performance and reliability in production environments.

Enterprise AI Platform

An enterprise AI platform is a unified infrastructure that orchestrates the entire lifecycle of AI development, deployment, and governance within large-scale organizational environments.

Enterprise Data Science Platform

An Enterprise Data Science Platform is a comprehensive suite of tools and infrastructure designed to support the end-to-end data science workflow in large organizations.

Feature Dataset

A feature is an individual measurable property or characteristic of a phenomenon being observed. Features are the input variables used in predictive modeling and statistical analysis.

Hybrid Cloud AI

An architecture combining private and public cloud resources to optimize AI model development, security, data sovereignty, and computational scalability.

Integrated Development Environment

An integrated development environment (IDE) combines essential tools for software development into a single application, enhancing productivity for data scientists and engineers.

Load Balancer

A component that distributes incoming requests across multiple backend servers or model endpoints to improve throughput, resilience, and cost efficiency.

MLOps Lifecycle

The MLOps lifecycle encompasses the end-to-end process of developing, deploying, and maintaining machine learning models in production, integrating software engineering best practices with data science workflows.

MLOps Platform

An MLOps platform integrates tools and processes to streamline the machine learning lifecycle, from development to deployment and monitoring.

Model Context Protocol (MCP)

Model Context Protocol (MCP) standardizes how AI applications connect with external data sources and tools.

Model Development

Model development in data science involves creating, training, and refining predictive models to extract insights from data and solve complex problems.

Model Registry

A model registry is a centralized repository for managing and versioning machine learning models throughout their lifecycle.

Multi-Cluster Orchestration

The automated management and coordination of workloads across separate server clusters to ensure high availability, scalability, and disaster recovery.

NCUA (National Credit Union Administration)

NCUA is the federal agency that supervises and regulates all federal credit unions and state-charted credit unions. NCUA exam expectations now include documented AI governance, access controls, and data protection.

PII and PHI (Personally Identifiable Information and Protected Health Information)

PII is Personally Identifiable Information and PHI is Protected Health Information. AI workflows that process either class need a defined boundary, access control, and an audit trail to stay compliant.

Retrieval-Augmented Generation (RAG)

Retrieval-Augmented Generation (RAG) connects generative AI models to your private data to deliver accurate, context-aware, and up-to-date responses.

Role-Based Access Control (RBAC)

Role-Based Access Control limits system access based on user roles, ensuring employees only see data necessary for their specific job functions.

SHAP

SHAP (SHapley Additive exPlanations) is a game theoretic approach to explain the output of any machine learning model.

Sovereign AI

Vector Database

A specialized database that indexes high-dimensional vector embeddings to enable efficient semantic search, recommendation systems, and context retrieval for AI.

Virtual Air-Gap

Virtual Air-Gap is a security architecture that logically isolates sensitive systems within a network without requiring physical disconnection.

Explore