Glossary J - P

KISS (Keep It Simple, Stupid) is a design principle that prioritises simplicity. Learn why the simplest solution is usually the best.

Open-source container orchestration platform originally developed by Google for automating deployment, scaling, and management of containerized applications.

Open-source framework for building LLM-powered applications with abstractions for chains, agents, memory, and tool integration.

Framework for building stateful multi-agent applications using directed graphs, part of the LangChain ecosystem.

Hierarchical multi-agent orchestration architecture where a central agent coordinates interactions between specialized agents, delegating subtasks, monitoring progress, and synthesizing results.

Large-scale neural network trained on vast text data to understand and generate natural language.

Use of Large Language Models (GPT-5, Claude Opus 4.5) as core of software development process to accelerate 5x-12x.

Network device or service that distributes incoming traffic across multiple servers to improve application availability, scalability, and fault tolerance.

Efficient fine-tuning technique that inserts small trainable matrices into a frozen model, drastically reducing the resources needed to adapt LLMs.

Metric estimating the total revenue a customer generates throughout their entire relationship with the business, key for determining acquisition investment.

Open standard by Anthropic that standardizes how AI agents connect to external tools, data sources, and services.

Architectural style that structures an application as a collection of small, independently deployable services communicating through APIs.

Technique for transferring knowledge from a large model (teacher) to a smaller one (student), reducing inference costs while preserving most of the performance.

Architectural style where all application functionality is deployed as a single executable unit.

Metric representing the normalized monthly recurring revenue of a subscription business, the fundamental basis for measuring growth and predicting revenue.

System that coordinates multiple specialized AI agents (QA, Backend, Frontend, Product) working together to complete complex software development tasks.

SaaS architecture pattern where a single application instance serves multiple customers (tenants) with data isolation and independent configuration.

Product with the minimum set of features needed to validate a business hypothesis with real users. Not a beta version, but a learning experiment.

Performance anti-pattern where an initial query triggers N additional queries, significantly degrading application performance.

Customer loyalty metric measuring the likelihood of users recommending a product or service, classifying respondents as promoters, passives, and detractors.

Practice of understanding system behavior through its external outputs: logs, metrics, and traces, enabling investigation of unknown problems beyond traditional monitoring.

Management framework defining ambitious, measurable objectives with quantifiable key results to align teams and track progress transparently.

Decentralized architecture where agents operate as equals without a central coordinator, communicating directly with each other and making collaborative decisions through emergent consensus.

Prioritized, dynamic list of all pending work on a product, managed by the product owner as the single source of truth for the development team.

Growth strategy where the product is the primary vehicle for customer acquisition, conversion, and expansion, reducing dependency on sales teams.

State where a product satisfies real market demand sustainably, validated by retention metrics, organic growth, and willingness to pay.

Errors affecting real users

Design of optimized instructions for LLMs. Critical skill for Agent-Ops engineers, reduces costs 60-80%.

Infrastructure-as-code platform that lets you define and manage cloud resources with general-purpose programming languages across any cloud provider.