AI/ML Glossary

Mechanism that allows models to selectively weight and focus on relevant parts of input, fundamental in Transformer architecture.

Deep learning architecture based on the attention mechanism, foundation of all modern LLMs and generative models.

Computational resources, cloud services, and technical systems required to develop, train, and deploy artificial intelligence models at scale.

Measurement frameworks and key performance indicators (KPIs) for assessing AI system effectiveness, business impact, and organizational value.

Field of research focused on ensuring that artificial intelligence systems operate safely and in alignment with human values.

Methodologies and frameworks for assessing AI system performance, reliability, safety, and alignment with requirements in non-deterministic environments.

AI model pre-trained on large-scale data that serves as a foundation for adaptation to multiple downstream tasks through fine-tuning or prompt-based learning.

AI models that generate new content (text, images, audio, code) based on patterns learned from training data and user instructions (prompts).

A deep learning model with billions of parameters, trained on massive amounts of text to understand and generate natural language.

Assessment and measurement of how AI models behave across different inputs, including reliability, consistency, and failure modes.

Discipline of computer science and AI that studies and develops methods for enabling computers to understand, interpret, and generate human natural language.

Security testing methodology where authorized teams simulate adversarial attacks to identify vulnerabilities and weaknesses in systems.

Process of segmenting text into discrete units (tokens) for processing by language models.

Vector representation of text that captures semantic relationships in high-dimensional space. Foundation of RAG and semantic search.

OpenAI's multimodal frontier language model, capable of processing and generating text and images with top-tier performance on academic benchmarks and real-world tasks.

Family of open-weights language models developed by Meta, available in sizes from 7B to 405B parameters, with frontier performance competitive with closed-source models.

Generative machine learning model for creating images from textual descriptions (prompts), based on latent diffusion models. Developed by Stability AI in 2022, it combines compression in latent space with iterative denoising processes.

Open-source platform and organization providing pre-trained NLP models, software libraries, and infrastructure for the machine learning community.

AI research organization that develops and distributes frontier language models (GPT) through commercial API, with dominant influence on contemporary landscape.

Output from an AI model that appears plausible but contains false, invented, or unsupported information.

Systematic examination of why artificial intelligence projects fail, including causes, patterns, and lessons for improving project success rates.

Process of implementing and deploying artificial intelligence systems across organizations for business value creation and operational efficiency.

Product with just enough features to attract early-adopter customers and validate a product idea with minimum effort.

Degree to which a product satisfies strong market demand, critical moment for sustainable startup growth.

Realization of a method or idea to demonstrate technical and commercial feasibility before major investments.

Prompting technique that enhances LLM reasoning by requiring explicit step-by-step reasoning before providing the final answer.

Process of adapting a pre-trained model to a specific task or domain through training on targeted data.

Discipline of designing and optimizing inputs to LLMs to obtain desired outputs.

Alignment technique that uses human feedback to train LLMs to produce useful, safe, and intent-aligned outputs.

Architectural pattern that combines information retrieval from a knowledge base with LLM generation for answers grounded in specific documents.

Optimization techniques and strategies for improving content visibility in AI-powered search engines and generative AI overviews.

Ability of a nation or organization to control and maintain ownership of cloud infrastructure and data, often a requirement for government and sensitive operations.

Attributes that allow an organization to outperform its competitors and generate superior value for customers and stakeholders.

Cost advantages that enterprises obtain due to size, output, or scale of operation.

Long-term plans and tactics used by nations and organizations to secure competitive advantage in global markets and technology development.

Percentage of total sales in an industry generated by a particular company, key indicator of competitive position.

Phenomenon where a product or service gains value as more people use it, creating self-reinforcing growth.

Practice of optimizing websites and content to rank higher in search engine results, including traditional SEO and AI-driven search optimization.

Systematic development and recruitment of skilled professionals at all career levels to meet current and future organizational needs.

Dependency on a vendor for products and services unable to switch without substantial costs.

Database optimized for storage and similarity search on high-dimensional data such as text and multimodal embeddings.

Framework of policies, regulations, and institutional structures that guide the development, deployment, and use of artificial intelligence systems.

Hypothetical type of artificial intelligence that matches or surpasses human cognitive abilities across a wide range of tasks, demonstrating flexible reasoning and learning comparable to human intelligence.

Phenomenon whereby the European Union's regulatory standards become global standards through market power and regulatory influence, shaping international business practices beyond EU borders.

European Union regulation on artificial intelligence establishing harmonized rules for the development, placement on market, and use of AI systems based on risk classification.

Adherence to laws, regulations, and standards set by governmental and regulatory bodies, including AI-specific regulations like the EU AI Act.

Process of improving one's skills, knowledge, and experience to advance professional capabilities and achieve career goals.

An organizational culture that values continuous, constructive, and bidirectional feedback as a tool for growth and improvement.

Analysis of the relationship between human skill development and AI automation, examining whether businesses invest in junior talent or replace with AI systems.

A recurring meeting between manager and direct report for feedback, professional development, and alignment on priorities and goals.

A shared belief that the team is a safe environment for interpersonal risk-taking, where one can speak up, make mistakes, and ask for help without fear of negative consequences.

A leadership philosophy where the leader's primary goal is to serve the team, putting their needs before their own.

The degree to which a team achieves its goals, meets its members' needs, and grows its capacity to work together over time. Hackman's research shows it depends mostly on structural conditions, not on individual traits.

Predictable revenue a company expects to receive annually from subscriptions or recurring contracts.

The rate at which a company spends its capital before generating positive cash flow, a critical metric for startups.

Total cost of acquiring a new customer, including marketing and sales expenses and all related costs.

Accounting distinction between capital expenditure (long-term assets) and operational expenditure (day-to-day recurring costs).

Rate at which customers stop subscribing to a service, critical metric for subscription businesses and retention.

Total revenue a business can expect from a single customer over their entire relationship.

Earnings Before Interest, Taxes, Depreciation, and Amortization: measure of a company's core operating profitability.

Predictable revenue a company expects to receive monthly from recurring subscriptions.

Financial ratio measuring profitability as a percentage of revenue, key indicator of operational efficiency.

Financial metric measuring the gain or loss generated on an investment relative to the amount invested.

Comprehensive assessment of all direct and indirect costs associated with an asset over its entire lifecycle.

Revenues and costs associated with a single unit of a business model, fundamental for evaluating scalability and profitability.

Professional activities performed in a state of distraction-free concentration that push cognitive capabilities to their limit, creating value and being difficult to replicate.

Time management tool that helps prioritize activities by distinguishing between urgency and importance.

A mental state of complete immersion in an activity, characterized by intense focus, loss of time perception, and intrinsic pleasure in task execution.

Personal productivity methodology that frees the mind from task tracking by transferring all commitments to a reliable external system.

The belief that one's abilities and intelligence can be developed through dedication, effort, and learning, in contrast to a fixed mindset that views them as innate traits.

Statistical principle stating that approximately 80% of effects come from 20% of causes, applicable to multiple domains.

Time management technique using 25-minute focused work intervals alternated with short breaks.

An iterative and incremental approach to software development that emphasizes collaboration, adaptability, and continuous delivery of value.

Systematic examination of source code by peers to identify bugs, improve quality, and share knowledge.

Process of improving content quality, structure, and presentation to meet the needs of users, search engines, and AI systems.

Cross-industry standard process for data mining and data science projects, structured in six iterative phases.

Process-oriented methodology combining Agile and DevOps practices to improve data analytics quality and reduce cycle times.

An iterative methodology for creative problem-solving centered on users, combining empathy, ideation, and rapid experimentation.

A set of cultural practices, processes, and tools that integrate software development and IT operations to accelerate delivery and improve reliability.

Japanese philosophy of continuous improvement through small incremental changes involving all people in the organization.

A lean method to manage and improve work through workflow visualization, work-in-progress limits, and continuous optimization.

Measurable metric used to evaluate the success of an organization, team, or project in achieving specific objectives.

Methodology to improve productive efficiency through systematic elimination of waste and continuous process optimization.

A collaborative goal-setting framework that aligns organizations through ambitious objectives and measurable key results.

An agile software development technique where two programmers work together at one workstation, with one writing code (driver) and the other reviewing (navigator).

Processes and practices for ensuring that artificial intelligence systems meet quality standards, performance metrics, and reliability requirements.

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

An agile ceremony where the team reflects on the just-completed work process to identify concrete and actionable improvements.

An agile framework for managing complex product development through iterative cycles called sprints, defined roles, and structured ceremonies.

Criteria for defining goals that are Specific, Measurable, Achievable, Relevant, and Time-bound.

The implied cost of future rework required when choosing a quick but limited solution instead of a more robust approach.

A software development approach where tests are written before production code, following the Red-Green-Refactor cycle.

Sequential approach to software development where each phase must be completed before starting the next one.

Technical systems, protocols, and standards that support the delivery of web content and services, including SEO metadata and AI crawling configuration.