28 terms every decision-maker, architect, and product lead should know before scoping an AI project.
A structured pipeline where an LLM makes decisions at defined branch points, but the overall flow is deterministic.
An LLM-driven system that plans, calls tools, and executes multi-step tasks autonomously.
The framework of policies, controls, and audits that govern how AI is built and used in an organization.
Prompting technique where the model is asked to reason step-by-step before answering.
The field of AI that enables machines to interpret and act on visual information from images and video.
The maximum number of tokens a model can process in a single request (input + output).
Running AI models on-device or on local hardware instead of in the cloud.
A numerical vector that captures the semantic meaning of a piece of content.
A neural network that converts text (or images, audio) into dense numeric vectors capturing meaning.
Continuing training of a pre-trained model on domain-specific data to specialize its behavior.
Structured way for an LLM to invoke external functions or APIs with typed arguments.
Structured constraints that keep LLM outputs safe, on-scope and correctly formatted.
When an LLM generates plausible-sounding but factually incorrect output.
Combining dense (semantic) retrieval with sparse (keyword / BM25) retrieval for better recall.
A neural network trained on massive text corpora to understand and generate human-like language.
A parameter-efficient fine-tuning technique that trains small adapter matrices instead of full model weights.
An open standard that lets LLMs discover and call external tools, data, and services in a uniform way.
DevOps practices adapted to the lifecycle of machine learning and AI systems.
Models that understand and generate across multiple input types — text, images, audio, and video.
Extracting machine-readable text and structure from images, scans, and PDFs.
The practice of designing model inputs to produce reliable, task-specific outputs.
An attack where malicious instructions hidden in inputs override or subvert the LLM's intended behaviour.
Compressing a model by reducing the numerical precision of its weights, cutting memory and speed cost.
A second-stage model that re-scores retrieved candidates to improve final ranking quality in RAG.
A pattern that grounds LLM answers in your private documents by retrieving relevant snippets at query time.
Training technique that aligns LLMs to human preferences using reinforcement learning on preference data.
A compact LLM (1B–15B parameters) tuned for cost, latency, and on-device deployment.
A database optimized for storing and searching high-dimensional embeddings by similarity.
Take our 3-minute AI Readiness assessment and get a tailored roadmap for your organization.