AI Glossary
Plain-language definitions for AI, ML, and cloud engineering terms. 442 entries, A to Z.
A
- Abstract Factory Pattern A creational design pattern that provides an interface for creating families of related or dependent …
- Abstraction A fundamental programming principle that hides complex implementation details behind simplified …
- Access Control Models Frameworks for controlling who can access resources, including DAC, MAC, RBAC, and ABAC.
- ACID Properties The four guarantees that database transactions provide to ensure data reliability: Atomicity, …
- Activation Function What activation functions are, how they enable neural networks to learn non-linear patterns, and …
- Active Learning Framework for intelligently selecting the most informative data points to label, reducing annotation …
- Activity Diagram A UML behavioral diagram for modeling workflows, business processes, and algorithms with support for …
- Adapter Pattern A structural design pattern that converts the interface of a class into another interface that …
- Adversarial Machine Learning How evasion, poisoning, and model extraction attacks threaten ML systems, and the defenses available …
- Agentic AI What makes AI agentic vs assistive, autonomous task execution, tool use, planning capabilities, and …
- Agentic RAG Retrieval-Augmented Generation systems in which the language model actively plans, queries, …
- Aggregate Root What an aggregate root is in DDD, how it enforces consistency boundaries, and how to design …
- AI Agent What AI agents are, how they autonomously plan and execute tasks, and the architectural patterns …
- AI Agents - Autonomous Task Execution What AI agents are, how they differ from simple LLM calls, the key design patterns, and what makes …
- AI Gateway A centralized proxy layer that routes, governs, monitors, and optimizes requests to LLM providers, …
- AI Guardrails - Safety and Compliance Controls What AI guardrails are, the types of controls they enforce, how to implement them in enterprise …
- AI Hardware Comparing GPUs, TPUs, and custom ASICs from NVIDIA, Google, Groq, and Cerebras for training and …
- AI Literacy What AI literacy means, why it matters for organizations adopting AI, and what competencies are …
- AI Red Team A dedicated adversarial testing team that probes AI systems for vulnerabilities, biases, safety …
- AI Safety What AI safety is, the categories of harm it addresses, and the technical and organizational …
- AI Watermarking How invisible signatures are embedded in AI-generated text, images, and audio to enable detection …
- AIOps What AIOps means, how AI-driven operations improve alerting, root cause analysis, and automated …
- Amazon Aurora What Aurora is, how it provides managed relational database performance, and when to choose Aurora …
- Amazon Bedrock AgentCore Amazon Bedrock AgentCore is a managed runtime and governance layer for deploying, operating, and …
- Amazon DynamoDB What DynamoDB is, how its key-value model works, and when to choose DynamoDB for AI application …
- Amazon Kinesis What Amazon Kinesis is, how it processes streaming data in real time, and when to use Kinesis versus …
- Anomaly Detection Methods for identifying outliers and unusual patterns in data, including Isolation Forest, One-Class …
- Apache Kafka What Kafka is, how it provides distributed event streaming, and when to choose Kafka for AI data …
- API - Application Programming Interface What an API is, REST vs GraphQL vs gRPC, authentication patterns, rate limiting, and how AI services …
- API Gateway What an API gateway is, how it manages API traffic, and when to use managed gateways versus custom …
- ArchiMate An open and independent modeling language for enterprise architecture, maintained by The Open Group.
- ARIMA Autoregressive Integrated Moving Average model for time series forecasting, including SARIMA for …
- Association Rule Mining Pattern discovery techniques including Apriori and FP-Growth algorithms for finding frequent …
- Asymmetric Encryption Public-key cryptography using mathematically related key pairs, including RSA and elliptic curve …
- Attention Mechanism What attention mechanisms are, how they enable transformers to process sequences, and why they …
- Authentication and Authorization (AuthN/AuthZ) The distinction between verifying identity (authentication) and granting access permissions …
- Auto-Scaling What auto-scaling is, how it adjusts capacity dynamically, and how to configure scaling policies for …
- Autoencoder What autoencoders are, how they learn compressed data representations, and practical applications in …
- Automata Theory and Formal Languages The study of abstract computational machines and the formal languages they recognize, forming the …
- Automated Decision-Making The legal framework under GDPR Article 22 governing decisions made solely by automated systems, …
B
- Backpropagation What backpropagation is, how it computes gradients for neural network training, and why it matters …
- Batch Normalization What batch normalization is, how it stabilizes neural network training, and when to apply it in …
- Bayesian Optimization Gaussian process-based sequential optimization method for efficient hyperparameter tuning of …
- Bias-Variance Tradeoff What the bias-variance tradeoff is, how it explains model generalization, and how to use it to guide …
- Binary and Number Systems in Computing How computers represent all data in base-2 (binary), why transistors make this fundamental, and how …
- Blue-Green Deployment What blue-green deployment is, how it works, why it matters for zero-downtime AI model updates, and …
- Boolean Algebra and Logic Gates The mathematical foundation of digital logic, from George Boole's algebraic system to physical …
- Bounded Context What a bounded context is, how it defines model boundaries in DDD, and how it guides microservice …
- BPMN - Business Process Model and Notation A standardized graphical notation for specifying business processes in workflow and process …
- Bridge Pattern A structural design pattern that decouples an abstraction from its implementation so that the two …
- Builder Pattern A creational design pattern that separates the construction of a complex object from its …
- Business Process Management (BPM) A discipline for designing, executing, monitoring, and optimizing organizational business processes.
C
- Canary Deployment What canary deployment is, how gradual traffic shifting works, which metrics to watch, and how to …
- CAP Theorem A fundamental theorem in distributed systems stating that a distributed data store can provide at …
- CDN - Content Delivery Network What CDNs do, how CloudFront accelerates content delivery, and when to use a CDN for AI application …
- CE Marking for AI The CE marking applied to high-risk AI systems under the EU AI Act, indicating conformity with EU …
- Chain of Responsibility Pattern A behavioral design pattern that passes a request along a chain of handlers, where each handler …
- Chain-of-Thought (CoT) Prompting Eliciting intermediate reasoning steps from language models to improve performance on multi-step …
- Change Data Capture What change data capture (CDC) is, how Debezium and AWS DMS enable real-time data replication, and …
- Chaos Engineering What chaos engineering is, how controlled experiments improve system resilience, and how to start …
- CI/CD - Continuous Integration and Continuous Delivery What CI/CD is, why it matters for AI projects, the tools involved, and the AI-specific …
- CIA Triad - Confidentiality, Integrity, Availability The three fundamental objectives of information security that guide the design and evaluation of …
- Circuit Breaker Pattern What the circuit breaker pattern is, why AI services need it for handling model timeouts and rate …
- Class Diagram The most widely used UML diagram type, showing classes with their attributes and methods along with …
- Clean Architecture What clean architecture is, how dependency inversion organizes code layers, and when this structure …
- Client-Server Architecture A distributed architecture where clients send requests to centralized servers that provide services …
- Cloud Governance The framework of policies, processes, and controls that organizations use to manage cloud resources, …
- Clustering What clustering is, major clustering algorithms, and practical applications for grouping data …
- CMMI - Capability Maturity Model Integration A process improvement framework that helps organizations improve performance across projects, …
- COBIT - Control Objectives for Information and Related Technologies An IT governance and management framework developed by ISACA for aligning IT with business goals.
- Code Smells and Refactoring Indicators of design problems in code and systematic techniques for improving code structure without …
- Command Pattern A behavioral design pattern that encapsulates a request as an object, allowing parameterization of …
- Compiler and Interpreter Programs that translate human-readable source code into machine-executable instructions, through …
- Complexity Classes Classifications of computational problems by resource requirements, including P, NP, and …
- Component Diagram A UML structural diagram that shows the organization of system components, their interfaces, and the …
- Component-Driven Development Component-driven development is the practice of building UIs from isolated, reusable components, …
- Composite Pattern A structural design pattern that composes objects into tree structures to represent part-whole …
- Composition Over Inheritance A design principle that favors object composition over class inheritance for code reuse, resulting …
- Compound AI System An AI architecture that combines multiple models, retrievers, tools, and programmatic logic to solve …
- Computer Vision What computer vision is, how it works in AI applications, and how AWS Rekognition, Azure Computer …
- Concept Drift What concept drift is, how the relationship between inputs and outputs changes over time, and …
- Concurrency and Synchronization Techniques for managing concurrent execution in operating systems and applications, including …
- Conformity Assessment The EU AI Act process for evaluating whether a high-risk AI system meets regulatory requirements …
- Confusion Matrix What a confusion matrix is, how to read it, and how it connects to precision, recall, and other …
- Container Registry What container registries are, how ECR, Docker Hub, Azure ACR, and GCP Artifact Registry compare, …
- Continuous Integration (CI) Fundamentals The practice of frequently merging code changes into a shared repository with automated builds and …
- Continuous Training What continuous training is, how automated retraining pipelines keep ML models current, and the …
- Contract Testing What contract testing is, how it verifies service integration agreements, and when to use it instead …
- Contrastive Learning How contrastive learning methods like SimCLR, CLIP, and MoCo learn useful representations by …
- Convolutional Neural Network How CNNs extract spatial features from images and why architectures like ResNet, EfficientNet, and …
- Cost Optimization (Well-Architected Pillar) The Well-Architected pillar covering right-sizing, reserved capacity, spot instances, and cost …
- CPU Scheduling Operating system algorithms that determine which process or thread runs on the CPU, including FCFS, …
- CQRS - Command Query Responsibility Segregation What CQRS is, how it separates read and write models, and when this pattern improves AI application …
- Critical Path Method (CPM) A schedule analysis technique that identifies the longest sequence of dependent activities …
- Cross-Validation What cross-validation is, how it provides robust model performance estimates, and when to use …
- Cybernetics The study of control, communication, and feedback in systems, whether mechanical, biological, or …
D
- Data Catalog What a data catalog is, how metadata management and data discovery tools help AI teams find, …
- Data Contract What data contracts are, how schema-first agreements between data producers and consumers prevent …
- Data Controller The entity that determines the purposes and means of processing personal data under GDPR, bearing …
- Data Drift What data drift is, how input data distributions change over time, and methods for detecting and …
- Data Lake What a data lake is, how it stores raw data at scale, and when to use a data lake versus a data …
- Data Lineage What data lineage is, how tracking data from origin through transformations supports compliance, …
- Data Mesh What data mesh is, how it decentralizes data ownership, and when this organizational pattern is …
- Data Modeling The process of creating visual representations of data structures at conceptual, logical, and …
- Data Processor An entity that processes personal data on behalf of a data controller under GDPR, relevant to AI …
- Data Product A self-contained, discoverable unit of data managed as a product with clear ownership, quality …
- Data Quality What data quality means for AI systems, the dimensions of data quality, and how validation, …
- Data Sovereignty The principle that data is subject to the laws and governance of the country or region where it is …
- Data Structures for AI Applications Arrays, hash maps, trees, graphs, queues, and vector stores - how the choice of data structure …
- Data Warehouse What a data warehouse is, how it supports analytical queries on structured data, and how it …
- Database Indexing Data structures that improve query performance by providing fast lookup paths to rows in database …
- Database Normalization A systematic approach to organizing relational database tables to reduce data redundancy and improve …
- Database Transactions Units of work in a database that group multiple operations into a single atomic, consistent, …
- DBSCAN Density-Based Spatial Clustering of Applications with Noise, a clustering algorithm that discovers …
- Deadlock A condition where two or more processes are permanently blocked, each waiting for a resource held by …
- Decision Tree What decision trees are, how they make predictions through hierarchical rules, and their role as …
- Decorator Pattern A structural design pattern that attaches additional responsibilities to an object dynamically, …
- Deep Learning What deep learning is, how it differs from traditional machine learning, and when deep learning is …
- Deep Reinforcement Learning How deep RL algorithms like DQN, PPO, and A3C combine neural networks with reward-based learning, …
- Dependency Inversion Principle (DIP) A SOLID design principle stating that high-level modules should not depend on low-level modules, and …
- Deployment Diagram A UML structural diagram that shows the physical deployment of software artifacts on hardware nodes, …
- Design Systems What design systems are, how Brad Frost's Atomic Design formalized the methodology, and how …
- Design Tokens What design tokens are, how Jina Anne coined the term at Salesforce in 2014, and how the W3C Design …
- DevSecOps What DevSecOps means, how it integrates security into every stage of CI/CD, and why shifting …
- Diffusion Models What diffusion models are, how they generate images and other media, and their role in enterprise AI …
- Digital Signatures and Certificates Public key infrastructure (PKI) mechanisms for verifying authenticity and integrity using X.509 …
- Dimensionality Reduction What dimensionality reduction is, common techniques including PCA and t-SNE, and when to reduce …
- Direct Preference Optimization (DPO) An alignment method that fine-tunes language models directly on preference data without training an …
- Divide and Conquer An algorithmic paradigm that recursively breaks a problem into smaller subproblems, solves them …
- DNS - Domain Name System The hierarchical, distributed naming system that translates human-readable domain names into IP …
- Docker What Docker is, how containers package applications, and best practices for containerizing AI …
- Document Extraction Definition of document extraction, the main techniques (OCR, NLP, template-based), AWS services used …
- Domain-Driven Design (DDD) What domain-driven design is, how it aligns software architecture with business domains, and when to …
- DORA - Digital Operational Resilience Act EU regulation requiring financial entities to ensure ICT resilience, covering risk management, …
- DPIA - Data Protection Impact Assessment A structured process required under GDPR Article 35 to identify and mitigate data protection risks …
- Dropout What dropout is, how it prevents overfitting in neural networks, and practical guidance on when and …
- DRY Principle - Don't Repeat Yourself A software development principle stating that every piece of knowledge must have a single, …
- Dynamic Programming An algorithmic technique that solves complex problems by breaking them into overlapping subproblems …
E
- Earned Value Management (EVM) A project performance measurement technique that integrates scope, schedule, and cost metrics to …
- Edge Computing What edge computing is, how it brings computation closer to data sources, and when edge deployment …
- Elastic Stack (ELK) What the Elastic Stack is, how Elasticsearch, Logstash, and Kibana work together, and when to use it …
- ELT - Extract, Load, Transform What ELT is, how it differs from ETL, and why modern data architectures favor loading raw data …
- Embeddings - Vector Representations for AI Search What embeddings are, how they enable semantic search, which embedding models to use, and how to …
- Emotion and CSS-in-JS The CSS-in-JS paradigm introduced by Christopher Chedeau (Vjeux) in 2014 and the Emotion library …
- Encapsulation A fundamental object-oriented programming principle that bundles data and the methods that operate …
- End-to-End Testing What end-to-end testing is, how browser automation validates full-stack AI applications, and why E2E …
- Ensemble Methods What ensemble methods are, how combining models improves predictions, and when to use bagging, …
- Enterprise Architecture Overview An overview of the enterprise architecture discipline, covering its purpose, frameworks, and role in …
- Entity-Relationship Model A conceptual data modeling technique that represents data as entities, attributes, and …
- EPC Diagram - Event-driven Process Chain A flowchart-based modeling notation for business processes, originating from the ARIS framework.
- Error Budget What an error budget is, how it balances reliability with feature velocity, and how to implement …
- Essential Entity (NIS2) What an essential entity is under the NIS2 Directive, which sectors are classified as essential, and …
- ETL - Extract, Transform, Load What ETL is, how it powers data pipelines, and how it compares to ELT for modern data architectures.
- Event Sourcing What event sourcing is, why it matters for AI audit trails and pipeline replay, its relationship to …
- Event-Driven Architecture for AI What event-driven architecture is, how S3 triggers, EventBridge, and Step Functions patterns enable …
- Experiment Tracking What experiment tracking is, why systematic logging of ML experiments is essential, and the tools …
F
- F1 Score What the F1 score measures, when to use it as a model evaluation metric, and its limitations.
- Facade Pattern A structural design pattern that provides a simplified interface to a complex subsystem, reducing …
- Factory Method Pattern A creational design pattern that defines an interface for creating objects but lets subclasses …
- Feature Branching What feature branching is, how it isolates development work, and the tradeoffs compared to …
- Feature Flags What feature flags are, how they enable safe AI model rollouts, A/B testing, and instant rollback - …
- Feature Store What a feature store is, how it serves as a centralized repository for ML features, and why it …
- Few-Shot Learning What few-shot learning is, how it enables models to generalize from minimal examples, and practical …
- File Systems The methods and data structures operating systems use to organize, store, and retrieve data on …
- Fine-Tuning vs Prompt Engineering vs RAG The three main approaches to customizing LLM behavior for specific use cases - when each is …
- Firewalls and Network Security Network security devices and techniques that control traffic flow between networks, including packet …
- Flaky Test What flaky tests are, why they are especially common in AI systems, and strategies for managing …
- Flash Attention How Flash Attention makes transformer self-attention memory-efficient by restructuring computation …
- Floating-Point Arithmetic and Model Precision IEEE 754, FP32, FP16, BF16, and INT8 - how number precision determines model size, inference speed, …
- Flyweight Pattern A structural design pattern that uses sharing to support large numbers of fine-grained objects …
- Foundation Models What foundation models are, how they differ from task-specific models, the major model families, and …
- Function Calling Structured tool invocation by language models: how the model emits typed function calls, how …
G
- GAN - Generative Adversarial Network What GANs are, how generator-discriminator training works, and where GANs remain relevant alongside …
- Gantt Chart A horizontal bar chart used to visualize project schedules, showing tasks, durations, dependencies, …
- GDPR - General Data Protection Regulation The EU's comprehensive data protection law governing how personal data is collected, processed, and …
- GitHub Pages What GitHub Pages is, how it provides static site hosting directly from Git repositories, and its …
- GitOps What GitOps is, how it uses Git as the single source of truth for infrastructure and deployments, …
- Golden Dataset What a golden dataset is, how it serves as a curated evaluation benchmark for measuring AI model …
- Gradient Boosting Ensemble learning method that builds models sequentially to correct previous errors, including …
- Gradient Descent What gradient descent is, how it optimizes neural networks, and the variants used in modern deep …
- Grafana What Grafana is, how it visualizes metrics and logs, and best practices for building operational …
- Graph Algorithms Algorithms for traversing and finding paths in graphs, including BFS, DFS, Dijkstra's, and A*.
- Graph Neural Network How GNNs process graph-structured data for node classification, link prediction, and graph-level …
- Greedy Algorithms Algorithms that make the locally optimal choice at each step, aiming for a globally optimal or …
- Ground Truth What ground truth is in machine learning, how verified correct labels are obtained, and why ground …
- gRPC What gRPC is, how Protocol Buffers and streaming RPCs work, and why gRPC is well-suited for …
H
- Hallucination What AI hallucination is, why language models generate plausible but incorrect information, and …
- Hardware Constraints for AI Systems CPU vs GPU, VRAM limits, memory bandwidth, and how hardware choices determine what AI models you can …
- Hash Tables Data structures that map keys to values using hash functions for near-constant-time lookup, …
- Hashing Algorithms One-way functions that produce fixed-size digests from arbitrary input, including SHA-256, MD5, and …
- Heaps and Priority Queues Tree-based data structures that efficiently support finding and extracting the minimum or maximum …
- Helm Chart What Helm charts are, how they package Kubernetes deployments, and best practices for managing …
- Hexagonal Architecture What hexagonal architecture is, how ports and adapters decouple business logic from infrastructure, …
- Hierarchical Clustering Agglomerative and divisive clustering methods that produce a tree-like hierarchy of clusters …
- Homomorphic Encryption How homomorphic encryption enables computation on encrypted data, allowing ML inference without …
- HTTP and HTTPS The foundational web protocols for transferring hypertext documents and resources, with HTTPS adding …
- Human-in-the-Loop (HITL) Definition, why it matters in AI systems, implementation patterns, and when it is legally or …
- Hybrid Cloud What hybrid cloud is, why it matters for AI workloads with data gravity and compliance constraints, …
- Hyperparameter Tuning What hyperparameter tuning is, the main strategies for finding optimal settings, and how to approach …
I
- Idempotency What idempotency means, how idempotency keys work for API endpoints, and why safe retry behaviour is …
- Imbalanced Data Strategies for handling skewed class distributions including SMOTE, undersampling, class weighting, …
- Immutable Infrastructure What immutable infrastructure means, how it replaces mutable servers with disposable instances, and …
- Inference - Running AI Models in Production What inference means in AI context, the key operational parameters that matter (latency, throughput, …
- Inference-Time Compute The practice of allocating additional computation during model inference to improve reasoning …
- Information Systems Overview An overview of the information systems discipline, covering types of IS, their role in …
- Infrastructure as Code (IaC) What Infrastructure as Code is, and how Terraform, AWS CDK, and CloudFormation compare for managing …
- Inheritance and Polymorphism Core object-oriented programming mechanisms: inheritance creates class hierarchies for code reuse, …
- Integration Testing What integration testing is, how it verifies component interactions, and where test boundaries …
- Interface Segregation Principle (ISP) A SOLID design principle stating that no client should be forced to depend on methods it does not …
- Interpreter Pattern A behavioral design pattern that defines a representation for a language's grammar and provides an …
- IP Addressing and Subnetting The system of numerical addresses used to identify devices on IP networks, including IPv4, IPv6, …
- ISO/IEC 42001 - AI Management System The international standard specifying requirements for establishing, implementing, and improving an …
- Istio What Istio is, how it implements a service mesh on Kubernetes, and when the operational overhead is …
- IT Governance Overview An overview of IT governance principles and frameworks for ensuring IT investments support business …
- IT Service Management (ITSM) The discipline of designing, delivering, managing, and improving IT services to meet business needs.
- Iterator Pattern A behavioral design pattern that provides a way to access elements of an aggregate object …
- ITIL - Information Technology Infrastructure Library A framework of best practices for IT service management, originally developed by the UK government.
K
- K-Means Clustering What K-means clustering is, how the algorithm works, and practical guidance for applying it to …
- K-Nearest Neighbors (KNN) Instance-based lazy learning algorithm that classifies data points by majority vote of their nearest …
- Kiro What Kiro is, how AWS's spec-driven AI IDE structures development through requirements, design, and …
- KISS Principle - Keep It Simple A design principle stating that systems work best when they are kept simple rather than made …
- Knowledge Base (AI) What an AI knowledge base is, how it differs from a traditional knowledge base, vector stores, and …
- Knowledge Distillation How teacher-student training compresses large models into smaller, faster ones while preserving most …
- Kolmogorov-Arnold Network How KANs replace fixed activation functions with learnable functions on edges, offering …
- Kubernetes What Kubernetes is, how it orchestrates containers at scale, and when to use EKS versus simpler …
L
- Lakehouse Architecture What a lakehouse is, how it combines data lake flexibility with warehouse performance, and practical …
- Law of Demeter A design guideline for developing software, particularly object-oriented programs, that promotes …
- Layered Architecture A software architecture pattern that organizes components into horizontal layers with strict …
- Linear Regression Foundational supervised learning algorithm for continuous prediction, including Ridge, Lasso, and …
- Linked Lists, Stacks, and Queues Fundamental linear data structures for organizing and accessing data sequentially.
- Liskov Substitution Principle (LSP) A SOLID design principle stating that objects of a supertype should be replaceable with objects of a …
- LLM - Large Language Model What large language models are, how they work at a high level, key characteristics, and what they …
- LLM Routing Architectures that direct each request to one of several available language models based on cost, …
- LLM-as-a-Judge Using a language model as an automated evaluator of another model's outputs: methodology, …
- LLMOps - LLM Operations The practices, tools, and infrastructure for deploying, monitoring, and managing large language …
- Load Balancer What load balancers do, the types available on AWS, and how to choose the right one for your …
- Logistic Regression Binary and multinomial classification algorithm using the sigmoid function and log-loss …
- Long-Context Model How modern architectures handle 100K to 1M+ token contexts through positional encoding advances, …
- Loss Function What loss functions are, how they guide model training, and which loss functions apply to common AI …
M
- Material UI (MUI) The open-source React component library implementing Google's Material Design system, one of the …
- Mediator Pattern A behavioral design pattern that defines an object that encapsulates how a set of objects interact, …
- Memento Pattern A behavioral design pattern that captures and externalizes an object's internal state so it can be …
- Memory Management Operating system techniques for managing physical and virtual memory, including paging, …
- Message Queue What message queues are, how they decouple services, and when to use SQS versus other messaging …
- Mixture of Agents How multi-LLM collaboration frameworks improve response quality by combining outputs from diverse …
- Mixture of Experts (MoE) A neural network architecture in which only a small subset of parameters is activated for each …
- MLOps - Machine Learning Operations What MLOps is, how it applies DevOps principles to machine learning, and the practices that enable …
- Mocking Test doubles for AI systems: mocks, stubs, fakes, and spies explained, with guidance on when to use …
- Model Calibration Techniques for producing reliable probability estimates from classifiers, including Platt scaling …
- Model Card What a model card is, why standardized ML model documentation matters, and what information a model …
- Model Cards - AI Transparency Documentation What model cards document, why they matter for AI governance, and how to create one.
- Model Context Protocol (MCP) An open protocol that standardises how language models connect to tools, data sources, and external …
- Model Drift What model drift is, how model performance degrades over time in production, and the monitoring and …
- Model Drift and Data Drift What drift is, the three types (data, concept, prediction), how to detect them using SageMaker Model …
- Model Lineage The complete provenance record of an AI model, tracking its training data, code, hyperparameters, …
- Model Registry What a model registry is, how it provides versioned storage and lifecycle management for trained ML …
- Monolithic Architecture A software architecture where all components are built and deployed as a single, self-contained …
- Monorepo What a monorepo is, how Google's approach was documented in the landmark 2016 ACM paper, and how …
- Multi-Agent Orchestration Multi-agent orchestration is the pattern of coordinating multiple specialized AI agents to …
- Multi-Agent Systems Definition, architecture patterns, and frameworks for multi-agent AI systems - and the signals that …
- Multimodal Model How models like GPT-4o and Gemini process text, images, audio, and video together within a unified …
- MVC - Model-View-Controller An architectural pattern that separates application concerns into model (data), view (presentation), …
- MVVM - Model-View-ViewModel An architectural pattern that uses data binding to connect the View to a ViewModel, enabling …
N
- Naive Bayes Probabilistic classification algorithm based on Bayes' theorem with strong independence assumptions, …
- NAT Gateway What NAT gateways do, how they enable private subnet internet access, and cost considerations for …
- Network Protocols Overview A survey of essential network protocols beyond TCP, UDP, and HTTP, including ARP, ICMP, DHCP, FTP, …
- Neural Architecture Search How automated methods discover optimal neural network architectures using reinforcement learning, …
- Neural Network What neural networks are, how they learn from data, and where they fit in modern AI system …
- Neural Radiance Field How NeRF reconstructs photorealistic 3D scenes from 2D images using neural networks to represent …
- Neuromorphic Computing How brain-inspired spiking neural networks and specialized hardware like Intel Loihi enable …
- Next.js The React framework created by Vercel (formerly Zeit) in 2016 that popularized server-side …
- NIS2 - Network and Information Security Directive The EU's updated cybersecurity directive requiring essential and important entities to implement …
- NIST AI RMF - AI Risk Management Framework The US National Institute of Standards and Technology's voluntary framework for managing risks in AI …
- Node.js The server-side JavaScript runtime created by Ryan Dahl in 2009, built on Chrome's V8 engine with an …
- NoSQL Databases A broad category of non-relational database systems designed for specific data models and access …
- npm The package manager for Node.js created by Isaac Schlueter in 2010, which established the registry …
- Number Systems and Encoding Binary, hexadecimal, and character encoding systems (ASCII, Unicode) that underpin how computers …
O
- OAuth OAuth is an open standard for delegated authorization, originating from Blaine Cook and Chris …
- Object-Oriented Programming (OOP) Classes, objects, inheritance, encapsulation, and polymorphism - how OOP concepts apply directly to …
- Observability What observability means, the three pillars of logs, metrics, and traces, and why AI systems need …
- Observer Pattern A behavioral design pattern that defines a one-to-many dependency between objects so that when one …
- Onion Architecture An architecture pattern placing the domain model at the core with infrastructure concerns on the …
- Online Learning Incremental machine learning approach that updates models continuously with streaming data rather …
- Open Practice Library What the Open Practice Library is, its key practices for AI projects, and how it structures …
- Open-Closed Principle (OCP) A SOLID design principle stating that software entities should be open for extension but closed for …
- OpenAPI What the OpenAPI Specification is, how schema-first API design works, and why code generation from …
- Operating System Fundamentals The core concepts of operating systems including the kernel, system calls, resource management, and …
- Operational Excellence (Well-Architected Pillar) The Well-Architected pillar covering runbooks, automation, observability, incident response, and …
- OSI Model The Open Systems Interconnection model, a seven-layer conceptual framework that standardizes how …
- Overfitting What overfitting is, how to detect it, and practical strategies to prevent models from memorizing …
P
- PACELC Theorem An extension of the CAP theorem that addresses the trade-off between latency and consistency even …
- Pagefind What Pagefind is, how it provides static search for static sites using WebAssembly-based indexing, …
- PCA - Principal Component Analysis What PCA is, how it identifies principal components, and when to use it for dimensionality reduction …
- Penetration Testing Authorized simulated attacks on systems to identify security vulnerabilities before malicious actors …
- Performance Efficiency (Well-Architected Pillar) The Well-Architected pillar covering compute selection, storage, database, and networking choices - …
- Pipe and Filter Architecture An architecture pattern where data flows through a sequence of independent processing components …
- Platform Engineering What platform engineering means, how internal developer platforms accelerate AI/ML teams, and why …
- Playwright Playwright browser automation framework: what it is, key features, and why it is well-suited for …
- PMBOK - Project Management Body of Knowledge A comprehensive standard published by PMI that defines project management processes, knowledge …
- Ports and Adapters What the ports and adapters pattern is, how it structures application boundaries, and its …
- Positional Encoding How transformers represent sequence order using sinusoidal, rotary (RoPE), and ALiBi positional …
- Precision and Recall What precision and recall measure, how to choose between them, and why the tradeoff matters for …
- PRINCE2 - Projects IN Controlled Environments A structured, process-based project management methodology originally developed by the UK …
- Process Mining A data-driven technique for discovering, monitoring, and improving business processes by extracting …
- Processes and Threads The fundamental units of execution in operating systems, covering process lifecycle, thread …
- Programmatic Video Programmatic video is the practice of generating video content from code and data rather than manual …
- Progressive Delivery What progressive delivery means, how feature flags, canary releases, and automated rollback combine …
- Progressive Web App (PWA) Progressive Web Apps are web applications that use service workers, manifests, and modern browser …
- Prometheus What Prometheus is, how it collects and stores metrics, and how it fits into cloud-native monitoring …
- Prompt Caching Server-side caching of attention key/value tensors for repeated prompt prefixes, reducing latency …
- Prompt Engineering What prompt engineering is, why it matters in enterprise AI applications, and the most effective …
- Prompt Injection An attack technique where malicious input manipulates an LLM into ignoring its instructions, …
- Property-Based Testing What property-based testing is, why it is ideal for AI systems that cannot be tested with …
- Prototype Pattern A creational design pattern that creates new objects by cloning an existing instance, avoiding the …
- Proxy Pattern A structural design pattern that provides a surrogate or placeholder for another object to control …
- Pruning How structured and unstructured pruning reduce neural network size by removing redundant weights, …
- Pub/Sub - Publish-Subscribe Pattern What the pub/sub pattern is, how it enables event-driven architectures, and when to use SNS versus …
Q
R
- RAG - Retrieval Augmented Generation What RAG is, how it works, when to use it, and the common implementation pitfalls that reduce …
- RAG Evaluation Methods and metrics for measuring the quality of Retrieval Augmented Generation systems, covering …
- Random Forest What random forests are, how they combine decision trees for robust predictions, and when they are …
- React The declarative, component-based JavaScript UI library created at Facebook in 2013 that introduced …
- React Router The standard client-side routing library for React, created by Ryan Florence and Michael Jackson in …
- Reasoning Models Language models post-trained to allocate substantial inference-time compute to internal reasoning …
- Recurrent Neural Network How RNNs, LSTMs, and GRUs process sequential data, the vanishing gradient problem, and where …
- Recursion and Backtracking Self-referential functions and systematic trial-and-error with pruning for exploring solution …
- Red Teaming What red teaming is in AI, how adversarial testing discovers vulnerabilities and failure modes …
- Redis What Redis is, how it provides in-memory data storage, and common use cases for caching and …
- Reinforcement Learning What reinforcement learning is, how agents learn from rewards, and where RL applies in enterprise AI …
- Relational Algebra The formal mathematical foundation for querying relational databases, defining operations like …
- Reliability (Well-Architected Pillar) The Well-Architected pillar covering fault tolerance, disaster recovery, health checks, and scaling …
- Remix The full-stack React framework created by Ryan Florence and Michael Jackson in 2021, built on web …
- Remotion Remotion is a React framework for creating videos programmatically, treating video as code and …
- Repository Pattern What the repository pattern is, how it abstracts data access, and when to use it in AI application …
- Requirements Analysis The process of eliciting, analyzing, documenting, and validating the needs and constraints that a …
- Responsible AI What responsible AI is, the principles of fairness, transparency, accountability, and safety that …
- Right to Explanation The right under GDPR Article 22 for individuals to obtain meaningful information about the logic …
- Risk Register A structured document for recording identified project risks, their analysis, response plans, and …
- ROC Curve What ROC curves and AUC measure, how to interpret them, and when to use ROC versus precision-recall …
- Routing and Switching The fundamental network operations of forwarding data at Layer 2 (switching via MAC addresses) and …
- RPA - Robotic Process Automation Software robots that automate repetitive, rule-based tasks by mimicking human interactions with …
S
- Saga Pattern What the saga pattern is, how it manages distributed transactions across microservices, and …
- Search Algorithms Algorithms for finding elements in data structures, including linear search, binary search, and …
- Secure Multi-Party Computation How multiple organizations can collaboratively train ML models or compute joint analytics without …
- Security (Well-Architected Pillar) The Well-Architected pillar covering IAM, encryption, network security, and detection - and how it …
- Security Threat Modeling Structured approaches for identifying and prioritizing security threats, including STRIDE, DREAD, …
- Semantic Versioning What semantic versioning is, how MAJOR.MINOR.PATCH communicates change impact, and best practices …
- Semi-Supervised Learning Machine learning approach that leverages both labeled and unlabeled data through label propagation, …
- Sequence Diagram A UML behavioral diagram that shows how objects interact through messages exchanged over time, with …
- Server-Side Rendering (SSR) The web rendering paradigm where HTML is generated on the server for each request, from its origins …
- Serverless Computing What serverless computing means, how Lambda, Fargate, and Step Functions fit AI workloads, and when …
- Service Mesh What a service mesh is, how it manages service-to-service communication, and when the complexity is …
- Service-Oriented Architecture (SOA) An architectural style that structures applications as a collection of loosely coupled, …
- Sessionize What Sessionize is, how it manages conference call-for-papers, speaker profiles, and schedule …
- SHAP and LIME Post-hoc explanation methods for interpreting predictions of black-box machine learning models.
- Shared Responsibility Model What the shared responsibility model is, how AWS, Azure, and GCP divide security duties, and special …
- Sidecar Pattern What the sidecar pattern is, how it extends service functionality without code changes, and common …
- Single Responsibility Principle (SRP) A SOLID design principle stating that a class should have only one reason to change, meaning it …
- Single-Page Application (SPA) The web application architecture where a single HTML page dynamically rewrites its content in the …
- Singleton Pattern A creational design pattern that ensures a class has only one instance and provides a global point …
- Site Reliability Engineering (SRE) What SRE is, how it applies software engineering to operations, and key SRE practices for AI …
- SLA, SLO, and SLI What SLAs, SLOs, and SLIs are, how they relate to each other, and how to define them for AI …
- Snapshot Testing What snapshot testing is, how it captures and compares output snapshots for regression detection, …
- Socio-Technical Systems An approach recognizing that organizational performance depends on the joint optimization of social …
- Software Configuration Management The discipline of tracking and controlling changes to software artifacts, rooted in military …
- Software Development Lifecycle (SDLC) The structured process of planning, creating, testing, and deploying software systems through …
- Software Testing Fundamentals Core concepts of software testing including testing levels, techniques, and principles for verifying …
- SOLID Principles Five foundational object-oriented design principles that promote maintainable, flexible, and …
- Spec-Driven Development Spec-driven development is the pattern of writing structured specifications before code, formalized …
- Speech-to-Text (STT) What speech-to-text technology is, how AWS Transcribe, Azure Speech, and GCP Speech-to-Text compare, …
- SQL Fundamentals Structured Query Language, the standard language for defining, manipulating, and querying data in …
- Stackbit The visual editing platform for Jamstack sites founded by Ohad Eder-Pressman in 2019, which …
- Stakeholder Analysis The process of identifying stakeholders, assessing their interests and influence, and developing …
- State Machine Diagram A UML behavioral diagram that models the states of an object and the transitions between them in …
- State Pattern A behavioral design pattern that allows an object to alter its behavior when its internal state …
- State Space Model How structured state space models like Mamba and S4 achieve linear-time sequence modeling as an …
- Static Site Generation (SSG) The approach of pre-rendering web pages to static HTML at build time, tracing from Jekyll (Tom …
- Stored Procedures and Triggers Server-side database programs that encapsulate reusable logic and automatically respond to data …
- Strategy Pattern A behavioral design pattern that defines a family of algorithms, encapsulates each one, and makes …
- Stream Processing What stream processing is, how Flink, Spark Streaming, and Kafka Streams enable real-time data …
- Structured Output Constraining a language model to emit output that conforms to a specified schema (JSON, regex, …
- Subnet What subnets are, how they segment VPC networks, and best practices for subnet architecture on AWS.
- Supervised Learning What supervised learning is, how it works with labeled data, and when to choose it over other …
- Supply Chain Security Cybersecurity practices for managing risks across the chain of vendors, open-source components, and …
- Support Vector Machine (SVM) Margin-maximizing classifier that uses the kernel trick to handle high-dimensional and non-linear …
- Sustainability (Well-Architected Pillar) The Well-Architected pillar added in 2021 covering efficient resource usage, managed services, and …
- Symmetric Encryption Encryption algorithms that use the same key for both encryption and decryption, including AES and …
- Synthetic Data What synthetic data is, how artificially generated data is used for ML training and testing, and the …
- Systems Theory An interdisciplinary framework for studying complex systems as wholes, focusing on interactions, …
T
- t-SNE Non-linear dimensionality reduction technique for visualizing high-dimensional data in two or three …
- TCP and UDP The two primary transport-layer protocols of the Internet: TCP provides reliable, ordered delivery …
- TCP/IP Model The four-layer Internet protocol suite that defines how data is transmitted across networks, forming …
- Technical Debt The accumulated cost of shortcuts, compromises, and deferred improvements in a software system that …
- Template Method Pattern A behavioral design pattern that defines the skeleton of an algorithm in a base class, letting …
- Temporal Convolutional Network How causal dilated convolutions provide an efficient alternative to RNNs for sequence modeling with …
- Test Fixture What test fixtures are, how they provide predefined data and state for reproducible tests, and …
- Test-Driven Development The TDD red-green-refactor cycle and how it applies to AI application development where outputs are …
- Text-to-Speech (TTS) What text-to-speech technology is, how AWS Polly, Azure Speech, and GCP Text-to-Speech compare, and …
- Time Complexity and Big-O Notation An introduction to Big-O notation and how it describes the asymptotic behavior of algorithms.
- TinyML How machine learning runs on microcontrollers and resource-constrained devices using TensorFlow Lite …
- TLS/SSL Transport Layer Security and its predecessor Secure Sockets Layer, cryptographic protocols that …
- TOGAF - The Open Group Architecture Framework A comprehensive framework for enterprise architecture development, providing methods and tools for …
- Toil What toil is in the SRE context, how to identify it, and strategies for reducing operational burden …
- Token Budget The maximum number of tokens allocated for an LLM request or workflow, used to control costs, …
- Tokenization in AI What tokens are, how different models tokenize text, why token count matters for cost and context …
- Tool Use (in Language Models) The capability of a language model to invoke external tools: APIs, code execution, retrieval, …
- Training-Serving Skew What training-serving skew is, how mismatches between training and serving environments degrade …
- Transfer Learning What transfer learning is, how pre-trained models reduce training costs, and when to fine-tune …
- Transformer Architecture What the transformer architecture is, how it differs from prior approaches, and why it dominates …
- Trees and Binary Search Trees Hierarchical data structures including BSTs, AVL trees, red-black trees, and B-trees for efficient …
- Trunk-Based Development What trunk-based development is, how it differs from long-lived branches, and why it accelerates …
- Twelve-Factor App What the twelve-factor methodology is, how it guides cloud-native application design, and which …
- TypeScript The statically typed superset of JavaScript created by Anders Hejlsberg at Microsoft in 2012, …
U
- UMAP Uniform Manifold Approximation and Projection for faster dimensionality reduction that preserves …
- UML Overview The Unified Modeling Language, a standardized visual notation for specifying, constructing, and …
- Underfitting What underfitting is, how to identify it, and strategies to improve model performance when the model …
- Unit of Work Pattern What the unit of work pattern is, how it coordinates database changes, and when to apply it in …
- Unit Testing What unit testing is, how isolation and test doubles work, and assertion patterns relevant to AI …
- Unsupervised Learning What unsupervised learning is, how it discovers patterns without labels, and practical enterprise …
- Use Case Diagram A UML behavioral diagram that captures system functionality from the user's perspective, showing …
V
- Variational Autoencoder How VAEs learn structured latent spaces for generation, interpolation, and representation learning.
- Vector Database What vector databases are, how they enable semantic search, popular options including Pinecone, …
- Version Control Fundamentals Core concepts of version control systems including branching, merging, and distributed workflows …
- Virtual DOM The Virtual DOM is an in-memory representation of the real DOM introduced by React in 2013, enabling …
- Virtualization Fundamentals The technology of creating virtual instances of computing resources, including hypervisor-based …
- Vision Transformer How Vision Transformers (ViT) apply the transformer architecture to image recognition by treating …
- Visitor Pattern A behavioral design pattern that lets you add new operations to existing object structures without …
- Vite The next-generation frontend build tool created by Evan You in 2020 that leverages native ES modules …
- VPC - Virtual Private Cloud What a VPC is, how it provides network isolation on AWS, and essential VPC design considerations for …
W
- Web Components Web Components are a set of W3C standards (Custom Elements, Shadow DOM, HTML Templates) for creating …
- Webhooks Webhooks are user-defined HTTP callbacks that deliver real-time event notifications between web …
- WebSocket What WebSockets are, how they enable real-time bidirectional communication, and why they are used …
- Work Breakdown Structure (WBS) A hierarchical decomposition of project scope into manageable deliverables and work packages.
- Workflow Engine Software that automates the execution of business processes by coordinating tasks, decisions, and …
- WSJF - Weighted Shortest Job First Definition, formula, and how to adapt WSJF for scoring and prioritizing AI use cases and backlog …
X
Z
- Zachman Framework A two-dimensional classification schema for organizing the descriptive representations of an …
- Zero Trust Architecture What zero trust means, how it replaces perimeter-based security, and why AI model serving and data …
- Zero-Shot Learning What zero-shot learning is, how models perform tasks without examples, and when zero-shot approaches …
Open source projects