AI & ML Glossary

AI ethics is the interdisciplinary field studying the moral implications of developing and deploying artificial intelligence systems.

AI governance encompasses the policies, frameworks, regulations, and organizational structures that guide the development, deployment, and use of AI systems.

AI safety is the research field focused on ensuring that AI systems behave safely and as intended, particularly as they become more capable.

An API gateway is a reverse proxy that sits between clients and backend services, providing a single entry point for API requests.

API Routes are server-side endpoints in Next.js that handle HTTP requests, enabling the frontend application to implement backend logic without a separate server.

APIRouter is a FastAPI class for organizing API endpoints into modular, reusable groups that can be included in the main application with a shared prefix, tags, and dependencies.

The App Router is Next.js's file-system-based routing architecture (introduced in Next.js 13) that uses React Server Components by default, supports nested layouts, loading states, error boundaries, and parallel routes.

ASGI (Asynchronous Server Gateway Interface) is the Python standard for building asynchronous web applications and servers, succeeding the synchronous WSGI standard.

JWT (JSON Web Token) authentication is a stateless authentication mechanism where the server issues a digitally signed token containing user identity claims after successful login.

Amazon Web Services (AWS) is the world's largest cloud computing platform, providing over 200 services including computing (EC2, Lambda), storage (S3), databases (RDS, DynamoDB), machine learning (SageMaker, Bedrock), and networking.

AWS Lambda is Amazon's serverless compute service that runs code in response to events without provisioning or managing servers.

Microsoft Azure is a comprehensive cloud platform that provides strong AI capabilities through its partnership with OpenAI.

Bias auditing is the systematic evaluation of AI systems for unfair or discriminatory behavior across different demographic groups, use cases, and input distributions.

Celery is a distributed task queue library for Python that enables asynchronous execution of long-running tasks outside of the main application process.

ChromaDB is an open-source embedding database designed for simplicity and ease of use in AI applications.

A CI/CD (Continuous Integration / Continuous Deployment) pipeline automates the process of building, testing, and deploying software changes.

A Client Component in Next.js is a React component that runs in the browser, enabling interactivity through event handlers, state management (useState, useEffect), and browser APIs.

Google Cloud Run is a managed serverless platform that runs containerized applications with automatic scaling, including scale-to-zero.

A container is a lightweight, standalone, executable package that includes an application and all its dependencies (runtime, libraries, system tools, configuration) isolated from the host system through OS-level virtualization.

Containerization is the process of packaging an application and all its dependencies (runtime, libraries, system tools, configuration) into a standardized, portable unit called a container.

Content Security Policy (CSP) is a browser security standard that defines which resources (scripts, styles, images, connections) a web page is allowed to load.

CORS (Cross-Origin Resource Sharing) is a browser security mechanism that controls which web origins can make requests to your API.

CORS (Cross-Origin Resource Sharing) middleware in FastAPI configures which web origins, HTTP methods, and headers are permitted for cross-origin API requests.

CrewAI is a framework for building multi-agent AI systems where autonomous agents collaborate as a 'crew' to accomplish complex tasks.

Data poisoning is an adversarial attack where an attacker intentionally corrupts training data to manipulate a model's behavior.

Database migration is the process of evolving a database schema over time through versioned, incremental change scripts that can be applied (and often rolled back) in order.

Dependency Injection in FastAPI is the framework's built-in system for declaring, resolving, and injecting shared resources and logic into route handlers using the Depends() function.

Depends is FastAPI's dependency injection function that declares dependencies for route handlers, enabling automatic resolution and injection of shared resources.

Differential privacy is a mathematical framework for quantifying and limiting the privacy risk of individual data points when sharing aggregate statistics or training machine learning models.

Docker is a containerization platform that packages applications and their dependencies into portable, isolated containers that run consistently across any environment.

Docker is the standard containerization platform for packaging applications and dependencies into portable, isolated containers.

Edge deployment refers to running AI models and applications on edge devices or edge servers located close to end users or data sources, rather than in centralized cloud data centers.

Edge runtime refers to lightweight JavaScript/WebAssembly execution environments deployed to edge locations (CDN nodes) around the world, enabling code to run geographically close to users with minimal latency.

FastAPI is a modern, high-performance Python web framework for building APIs, built on top of Starlette (for async HTTP) and Pydantic (for data validation).

FastAPI is a modern, high-performance Python web framework for building APIs, leveraging Python type hints and Pydantic for automatic request validation, serialization, and OpenAPI documentation generation.

Google Cloud Platform (GCP) is Google's suite of cloud computing services, with particular strength in AI/ML through services like Vertex AI (MLOps platform), Gemini API, Cloud TPUs (custom AI accelerators), BigQuery (serverless data warehouse), and Cloud Run (serverless containers).

A health check is an endpoint or probe that reports whether a service is running and ready to accept requests.

Helm is the package manager for Kubernetes, using 'charts' (templated YAML configurations) to define, version, and deploy complex multi-component applications.

HIPAA (Health Insurance Portability and Accountability Act) compliance in AI refers to meeting the regulatory requirements for handling protected health information (PHI) when building AI systems for healthcare.

HTTPException is FastAPI's mechanism for returning HTTP error responses with specific status codes and detail messages.

Hybrid cloud is an IT architecture that combines on-premises infrastructure or private cloud with public cloud services, allowing workloads and data to move between them.

Image Optimization in Next.js is the automatic processing of images to serve the optimal format (WebP, AVIF), size, and quality for each user's device and viewport.

An inference pipeline is a production system that orchestrates the complete flow of processing an AI prediction request — from receiving input, through preprocessing, model inference, postprocessing, and returning results.

Infrastructure as Code (IaC) is the practice of managing and provisioning computing infrastructure through machine-readable configuration files rather than manual processes.

JSON Schema is a standard for describing the structure, constraints, and validation rules of JSON data.

Jupyter Notebook is an open-source interactive computing environment that combines executable code, rich text, equations, and visualizations in a single document (notebook).

Kubernetes (K8s) is an open-source container orchestration platform for automating deployment, scaling, and management of containerized applications.

Kubernetes (K8s) is the industry-standard container orchestration platform for deploying, scaling, and managing containerized applications in production.

LangChain is a popular open-source framework for building applications with large language models.

LangGraph is a library built on top of LangChain for constructing stateful, multi-agent AI applications using graph-based workflows.

LlamaIndex (formerly GPT Index) is a data framework for building LLM-powered applications that connect to external data sources.

Load balancing distributes incoming network traffic across multiple backend server instances to ensure no single server becomes a bottleneck, improving application availability and responsiveness.

Logfire is an observability platform by Pydantic that provides deep visibility into Python applications, with first-class support for AI/LLM workloads.

Metadata in Next.js refers to the system for defining HTML metadata (title, description, Open Graph tags, icons, robots directives) for pages using either a static metadata object or a dynamic generateMetadata function.

Middleware is a software layer that intercepts and processes HTTP requests and responses as they flow between the client and the application's route handlers.

Next.js Middleware is a function that runs before every request is processed, executing on the Edge Runtime at the CDN level.

Milvus is an open-source, cloud-native vector database built for scalable similarity search and AI applications.

Next.js is a React framework for building full-stack web applications with features like server-side rendering (SSR), static site generation (SSG), API routes, and edge functions.

Next.js is a React-based full-stack framework developed by Vercel that provides server-side rendering (SSR), static site generation (SSG), API routes, and edge functions out of the box.

NumPy (Numerical Python) is the foundational Python library for numerical computing, providing efficient multi-dimensional array operations and mathematical functions.

OAuth 2.0 is an authorization framework that enables third-party applications to access resources on behalf of a user without exposing their credentials.

OpenAPI Schema (formerly Swagger) is a standard specification for describing REST APIs in a machine-readable JSON/YAML format.

OpenTelemetry (OTel) is a vendor-neutral, open-source observability framework for generating, collecting, and exporting telemetry data — traces, metrics, and logs — from applications.

Pandas is the foundational Python library for data manipulation and analysis, providing the DataFrame data structure for working with tabular data.

A path parameter is a variable segment in a URL path that is extracted and passed to the route handler function — for example, /agents/{agent_id} extracts the agent_id value from the URL.

pgvector is a PostgreSQL extension that adds vector similarity search capabilities to PostgreSQL, enabling it to function as a vector database within your existing relational database.

Pinecone is a fully managed, cloud-native vector database service that handles infrastructure operations like scaling, replication, and backups automatically.

PostgreSQL is an advanced open-source relational database known for its reliability, feature richness, and extensibility.

Privacy-Enhancing Technologies (PETs) are a collection of techniques and tools designed to protect personal data while still enabling useful computation and AI model training.

A private cloud is a cloud computing environment dedicated exclusively to a single organization, either hosted on-premises or by a third-party provider.

Pydantic is the most widely used data validation library for Python, using type annotations to define data schemas and automatically validate, serialize, and transform data.

Pydantic AI is a Python agent framework built by the creators of Pydantic, designed for building production-grade AI agent applications with type safety and validation at the core.

Python's async/await syntax enables cooperative multitasking through coroutines — functions that can pause execution (await) while waiting for I/O operations (network calls, file reads, database queries) and resume when the result is ready.

PyTorch is the dominant open-source deep learning framework, developed by Meta AI, used for research and production training of neural networks including large language models.

Qdrant is an open-source vector database and similarity search engine written in Rust, designed for high performance and production reliability.

A query parameter is a key-value pair appended to a URL after the '?' character (e.g., /search?q=vector+database&limit=10) used to filter, sort, paginate, or configure API responses.

Rate limiting is a technique for controlling the frequency of requests a client can make to an API or service within a specified time window.

React Server Components (RSC) is a React architecture where components render on the server by default, sending only HTML and a minimal serialized component tree to the client.

Redis is an open-source, in-memory data structure store used as a database, cache, message broker, and streaming engine.

In FastAPI, response_model is a parameter on route decorators that specifies a Pydantic model defining the shape and types of the API response.

Responsible AI is an organizational approach to developing and deploying AI systems that are fair, transparent, accountable, safe, and privacy-preserving.

A reverse proxy is a server that sits in front of backend services and forwards client requests to the appropriate backend, adding capabilities like SSL termination, load balancing, caching, and request routing.

Route Handlers in Next.js are API endpoints defined using Web API Request and Response objects in route.ts files within the app/ directory.

Scikit-learn is the most widely used Python library for traditional machine learning, providing consistent APIs for classification, regression, clustering, dimensionality reduction, and model evaluation.

An SDK (Software Development Kit) is a collection of tools, libraries, documentation, and code samples that simplifies integration with a service or platform.

Server Actions are async functions in Next.js that execute on the server but can be invoked directly from Client Components, eliminating the need to manually create API endpoints for form submissions and data mutations.

A React Server Component (RSC) is a component that renders exclusively on the server, with its output (HTML) streamed to the client without sending its JavaScript to the browser.

Server-Sent Events (SSE) is a standard HTTP-based protocol for server-to-client streaming, where the server pushes events over a long-lived HTTP connection.

Serverless computing is a cloud execution model where the cloud provider dynamically manages server allocation, scaling, and infrastructure — developers deploy functions that run on demand without provisioning servers.

A serverless function is a single-purpose piece of code deployed to a serverless platform that executes in response to events, automatically scaling from zero with no infrastructure management.

HTTP status codes are three-digit numbers returned by a server indicating the result of an HTTP request.

Frontend streaming in Next.js and React refers to progressively sending rendered HTML and data to the client as it becomes available, rather than waiting for the entire page to be ready.

TensorFlow is an open-source deep learning framework developed by Google that provides tools for training and deploying machine learning models.

Terraform is an open-source infrastructure-as-code tool by HashiCorp that enables declarative provisioning and management of cloud resources across multiple providers (AWS, Azure, GCP) using HCL (HashiCorp Configuration Language).

TypedDict is a Python typing construct that defines dictionaries with a fixed set of string keys, each associated with a specific value type.

Uvicorn is a lightning-fast ASGI (Asynchronous Server Gateway Interface) server implementation for Python, built on uvloop and httptools for maximum performance.

Vercel is a cloud platform optimized for deploying frontend applications, particularly Next.js, with automatic CI/CD, global CDN, serverless functions, and edge computing.

A webhook is an HTTP callback mechanism where a server sends an automated POST request to a client-specified URL when a specific event occurs.

WebSocket streaming uses the WebSocket protocol to establish a persistent, full-duplex communication channel between the client and server, enabling real-time bidirectional data flow.

Zero data retention (ZDR) is a data handling policy where an AI service provider commits to not storing, logging, or using customer input data and model outputs after the API request is completed.