AI Prompting Glossary

Security measures that regulate who can view or use resources in a computing environment.

Iterative development methodology emphasizing collaboration, flexibility, and rapid delivery of working software.

A collaborative development approach where developers work alongside AI assistants to write, review, and improve code in real-time.

The discipline focused on ensuring AI systems behave safely, reliably, and according to human values and intentions.

Automated notification systems that inform teams of system issues, errors, or performance problems.

The process of creating well-structured, consistent, and user-friendly application programming interfaces.

A server-side architectural pattern that acts as a single entry point for all client requests to a system's backend services, intercepting API calls and performing functions like routing, authentication, and response aggregation.

The practice of using AI to generate code that integrates with external APIs, including authentication, data parsing, and error handling.

Practices and measures to protect application programming interfaces from attacks and unauthorized access.

The process of verifying the identity of users or systems before granting access to resources.

The process of determining what actions an authenticated user or system is permitted to perform.

Using software tools to execute tests automatically without manual intervention, improving efficiency and consistency.

A core feature of cloud computing that enables the dynamic and automatic adjustment of computational resources allocated to an application based on real-time demand, as measured by various performance metrics.

Development approach that focuses on the behavior of the application from the user's perspective, using natural language specifications.

The process of identifying unfair or discriminatory patterns in AI model decisions and outputs.

Deployment strategy using two identical production environments to enable zero-downtime releases and easy rollbacks.

Deployment strategy that releases changes to a small subset of users to detect issues before full rollout.

A prompting technique that encourages AI to show its reasoning process step-by-step before providing the final answer.

Design pattern that prevents cascading failures by temporarily stopping requests to failing services.

A metric measuring the percentage of code that is executed during testing, indicating test thoroughness.

The process of using AI to automatically create source code based on natural language descriptions or requirements.

The systematic examination of code by developers to find defects, improve quality, and share knowledge.

Using AI and automated tools to analyze code changes for quality, security, and compliance issues before human review.

Code that works correctly but indicates deeper design problems or areas that need refactoring.

Errors detected during code compilation before the program can be executed, typically syntax or type errors.

Using AI to create reusable UI components for web or mobile applications, including props, styling, and functionality.

A lightweight form of virtualization that bundles an application's code along with all of its necessary files, libraries, and runtime dependencies into a single, portable, and executable package known as a container.

A globally distributed network of proxy servers and data centers designed to deliver internet content including web pages, images, videos, and software to users more quickly, cheaply, reliably, and securely by reducing physical distance.

Techniques for reducing the amount of context while preserving essential information to fit within token limits.

The systematic approach to structuring and managing information provided to AI coding assistants to maximize their effectiveness and accuracy.

The maximum amount of text (measured in tokens) that an AI model can process and remember in a single conversation or request.

Practice of automatically deploying code changes to production after passing automated tests and quality checks.

Development practice where code changes are automatically tested and integrated into the main codebase frequently.

Techniques for generating code that works across multiple platforms (iOS, Android, Web) using frameworks like React Native or Flutter.

An attack that tricks users into executing unwanted actions on web applications where they're authenticated.

A security vulnerability where malicious scripts are injected into trusted websites, executed in users' browsers without their knowledge.

The process of encoding sensitive data to protect it from unauthorized access during storage and transmission.

A centralized repository designed to store, process, and secure massive amounts of data in its native, raw format. It can ingest and store all types of data without any upfront transformation or schema definition.

A centralized system that aggregates and consolidates large amounts of data from multiple sources into a single, consistent, and structured data store, specifically designed and optimized for business intelligence and analytics.

Code that exists in the codebase but is never executed or used, often left over from previous implementations.

The systematic process of finding, analyzing, and fixing bugs or errors in computer programs.

A specialized subfield of machine learning that employs artificial neural networks with multiple layers to learn intricate patterns and representations from large volumes of data. Its methodology is inspired by the structure and function of the human brain.

Design pattern that provides dependencies to an object rather than having the object create them itself.

Reusable solutions to commonly occurring problems in software design and architecture.

Cultural and technical practices that combine software development and IT operations to shorten development cycles.

A malicious cyberattack designed to render a target website, server, or network unavailable to legitimate users by overwhelming it with a massive flood of malicious internet traffic from multiple distributed sources.

A development approach where comprehensive documentation is created before implementation to guide the development process.

The process of converting readable data into an encoded format that can only be decoded with the proper key.

Testing complete user workflows from start to finish to ensure the entire system works as expected.

The practice of recording error information and system events for debugging and monitoring purposes.

Automated systems for detecting, tracking, and alerting on application errors in production environments.

Two distinct approaches to data integration. ETL transforms data before loading it into the target system, while ELT loads raw data first and then transforms it within the target system using its computational power.

A software design pattern built around the production, detection, consumption of, and reaction to events. Services are loosely coupled and communicate asynchronously through events rather than through direct, synchronous requests.

Programming construct for catching and managing runtime errors to prevent application crashes.

AI systems designed to provide clear explanations for their decisions and reasoning processes to users.

Design principle where systems detect and report errors immediately rather than continuing with invalid state.

A prompting technique where you provide a few examples of the desired output to guide the AI's response pattern.

The process of training an AI model on specific data to improve its performance for particular tasks or domains.

Adherence to the General Data Protection Regulation requirements for handling personal data of EU residents.

A broad category of artificial intelligence that is defined by its ability to create new, original content and ideas. Unlike traditional AI systems that recognize patterns or make predictions, generative models can produce text, images, videos, music, and software code.

Design approach where applications continue functioning with reduced capability when components fail.

A modern query language for APIs and a server-side runtime that allows clients to request exactly the data they need in a single API call, providing a more efficient and flexible alternative to traditional REST APIs.

When an AI model generates information that appears plausible but is factually incorrect or doesn't exist in the training data.

A one-way cryptographic function that converts input data into a fixed-size string, commonly used for password storage.

Automated tests that verify system components are functioning correctly and are available for use.

The ability of AI models to learn and adapt to new tasks based on examples or instructions provided within the conversation context.

The process by which AI models generate responses, predictions, or outputs based on input data and learned patterns.

The practice of managing and provisioning computing infrastructure through machine-readable definition files rather than through manual configuration or interactive graphical tools.

The process of checking user input to ensure it meets expected criteria and doesn't contain malicious content.

Tests that verify different components or services work together correctly as a group.

A development approach where you refine and improve AI-generated code through multiple rounds of feedback and regeneration.

A flexible framework used to implement Agile principles that emphasizes visualizing workflow, limiting work in progress, and actively managing the flow of tasks to improve delivery speed, efficiency, and predictability.

A type of very large deep learning model that has been pre-trained on immense quantities of text data. These models are designed to understand, summarize, generate, and interact with human language in a coherent and contextually relevant manner.

Technique for distributing incoming requests across multiple servers to optimize resource utilization and prevent overload.

Errors in program logic that cause incorrect results while the code runs without crashing.

A subset of AI that enables systems to automatically learn and improve from experience without being explicitly programmed.

The ease with which software can be modified, extended, and debugged over time.

A cyberattack where a perpetrator secretly positions themselves between two communicating parties to intercept, eavesdrop on, and potentially alter communication without either party being aware of their presence.

Model Context Protocol tools that enable AI assistants to interact with external systems and data sources for enhanced functionality.

Strategies for managing what information an AI assistant remembers across conversations and how to optimize context usage.

Architectural approach that structures applications as collections of loosely coupled, independently deployable services.

Testing technique using fake objects or services to simulate dependencies and isolate the code being tested.

The degradation of AI model performance over time as real-world data diverges from the training data.

Traditional architectural approach where all components of an application are interconnected and deployed as a single unit.

AI architectures that employ multiple specialized AI agents working together to accomplish complex development tasks.

Techniques for generating platform-specific mobile code for iOS (Swift/Objective-C) or Android (Kotlin/Java).

AI architectures inspired by biological neural networks, consisting of interconnected nodes that process information.

An open industry standard for authorization that enables third-party applications to obtain limited, delegated access to a user's data on another service without ever needing to handle the user's password.

The practice of instrumenting a software system to allow its internal state and behavior to be understood purely from its external outputs. It enables engineers to ask arbitrary questions about system performance and diagnose unforeseen problems.

A prompting technique where you provide a single example to demonstrate the desired output format or style.

Collaborative programming technique where two developers work together on the same code at the same workstation.

Simulated cyber attacks performed on systems to evaluate security and identify vulnerabilities before malicious actors do.

Continuous tracking of application performance metrics to identify bottlenecks and optimization opportunities.

The process of improving software performance by making code run faster, use less memory, or reduce resource consumption.

The art and science of crafting effective instructions and queries to get optimal responses from AI models.

Security attack where malicious inputs are crafted to manipulate AI models into producing unintended or harmful outputs.

Pre-designed prompt structures that can be reused and customized for specific development tasks or patterns.

Quantitative measures used to assess various aspects of code quality such as complexity, maintainability, and reliability.

A technique that combines AI generation with information retrieval from external knowledge bases to provide more accurate and up-to-date responses.

The process of restructuring existing code without changing its external behavior to improve readability and maintainability.

A sophisticated machine learning technique that integrates human judgment directly into the training process to optimize a model's behavior. Its primary goal is to align the model's outputs more closely with human goals, preferences, and values.

Techniques for generating CSS and component code that adapts to different screen sizes and devices.

The process of reverting to a previous stable version of software when issues are detected in the current version.

Errors that occur during program execution, after successful compilation, often due to invalid operations or unexpected conditions.

The process of cleaning user input by removing or escaping potentially dangerous characters and code.

The ability of a system to handle increased load by adding resources without compromising performance.

The gradual expansion of project requirements beyond the original specifications, often leading to increased complexity and delays.

Secure storage, distribution, and rotation of sensitive information like API keys, passwords, and certificates.

A comprehensive examination of systems, applications, and processes to evaluate security posture and compliance.

HTTP response headers that instruct browsers to implement additional security measures for web applications.

A search technique that understands the meaning and context of queries to find relevant information, often used in RAG systems.

A cloud-native software design paradigm that enables developers to build and run applications without having to provision or manage any of the underlying infrastructure. The cloud provider automatically manages the servers, operating systems, and scaling.

A hierarchical framework used in SRE to define, measure, and manage service reliability. SLIs measure performance, SLOs set internal targets, and SLAs define formal customer contracts with consequences.

A centralized, authoritative documentation system that serves as the primary reference for AI coding assistants.

Five fundamental design principles for object-oriented programming and software design that help developers create software systems that are more understandable, flexible, maintainable, and scalable.

A code injection technique where malicious SQL statements are inserted into application queries, potentially exposing or manipulating database data.

Secure communication protocols that encrypt data transmitted between clients and servers over the internet.

A detailed report showing the sequence of function calls that led to an error, used for debugging.

Techniques for generating code that manages application state using various patterns and libraries.

Automated examination of code without executing it to find potential bugs, security issues, and quality problems.

Errors that occur when code violates the grammatical rules of the programming language.

The high-level structure and organization of software systems, defining components and their relationships.

Accumulated shortcuts and compromises in code that need future improvement, created by quick fixes or suboptimal solutions.

A development methodology where tests are written before implementing the actual functionality.

The basic unit of text that AI models process, roughly equivalent to 3/4 of a word in English.

The process of teaching AI models to recognize patterns and make predictions by exposing them to large amounts of data.

Specialized prompting techniques for generating user interface designs and user experience patterns.

Testing individual components or functions in isolation to ensure they work correctly on their own.

Testing performed to determine whether a system meets business requirements and is ready for deployment.

Numerical representations of text that capture semantic meaning, used in AI systems for similarity matching and retrieval.

Using AI to generate commit messages, pull request descriptions, and manage code versioning workflows.

A software-based, digital emulation of a physical computer that operates on a physical host machine but functions as a completely separate and self-contained guest computer with its own virtualized hardware components.

A systematic review of systems and applications to identify security weaknesses and potential attack vectors.

A computer communications technology that enables a persistent, full-duplex communication channel over a single TCP connection. Unlike HTTP, WebSockets allow both client and server to send messages at any time, making it ideal for real-time web applications.

The use of technology to automate repetitive development processes and tasks to improve efficiency and reduce human error.

A modern security framework architected around the core principle of 'never trust, always verify.' It mandates stringent identity verification and authorization for every user and device attempting to access any resource, regardless of their location.

A prompting technique where you ask the AI to perform a task without providing any examples, relying on its pre-trained knowledge.