A reference from Weave

Engineering & AI glossary

Understand the metrics, models, and methods behind modern engineering. Clear definitions, practical examples, and a closer look at what the numbers actually mean.

All terms

2,008 terms

  • Instrumentation gap

    Instrumentation gap marks an unobserved or unreliable workflow measurement point.

    Engineering analytics

  • INT4 inference

    INT4 inference is the serving concept concerned with int4 inference during AI inference.

    Inference performance

  • INT8 inference

    INT8 inference is the serving concept concerned with int8 inference during AI inference.

    Inference performance

  • Intake capacity

    Intake capacity is a concept used in software delivery planning to describe a condition, relationship, estimate, or decision about engineering work.

    Flow and capacity planning

  • Integration debt

    Integration debt is a software maintenance concern describing a condition that can make future changes, verification, operation, or ownership harder. Its practical importance depends on supported behavior, rate of change, and the consequences of delay.

    Code quality and technical debt

  • Integration testing

    Integration testing checks whether two or more software components work together through their real or representative interfaces. It focuses on interactions such as application-to-database queries, service calls, message publishing, or framework configuration.

    Code quality and technical debt

  • Intention to treat analysis

    Intention to treat analysis is a statistical or measurement concept used to describe, compare, or interpret engineering data. Its meaning depends on the unit of analysis, data-generating process, and question being asked.

    Measurement and experimentation

  • INTer token latency

    INTer token latency is the serving concept concerned with inter token latency during AI inference.

    Inference performance

  • Inter-rater reliability

    Inter-rater reliability is a defined lens for examining AI system behavior with a stated task, evidence, and interpretation rule.

    Evaluations and benchmarks

  • Interface segregation

    A design principle that favors focused interfaces so clients depend only on what they use.

    Code quality and technical debt

  • Interface testing

    Interface testing is a software testing or test-design practice used to gather evidence about a defined risk, behavior, boundary, or operating condition. It makes the question under test explicit, identifies the inputs and observations that matter, and gives a team a repeatable basis for deciding whether the result is acceptable.

    Code quality and technical debt

  • Interference

    Interference is a statistical or measurement concept used to describe, compare, or interpret engineering data. Its meaning depends on the unit of analysis, data-generating process, and question being asked.

    Measurement and experimentation

  • Internal benchmark

    Internal benchmark is an analytical concept for using reference values to understand engineering performance, variation, or capability.

    Engineering analytics

  • Internal developer platform

    Internal developer platform is a way to organize, support, or evaluate software work so that teams can make useful progress with less avoidable friction. It is most valuable when connected to a concrete outcome and the local conditions of the team using it.

    Developer productivity

  • Internal validity

    Internal validity is a statistical or measurement concept used to describe, compare, or interpret engineering data. Its meaning depends on the unit of analysis, data-generating process, and question being asked.

    Measurement and experimentation

  • Internationalization testing

    Internationalization testing is a software testing or test-design practice used to gather evidence about a defined risk, behavior, boundary, or operating condition. It makes the question under test explicit, identifies the inputs and observations that matter, and gives a team a repeatable basis for deciding whether the result is acceptable.

    Code quality and technical debt

  • Interquartile range

    Interquartile range is a statistical or measurement concept used to describe, compare, or interpret engineering data. Its meaning depends on the unit of analysis, data-generating process, and question being asked.

    Measurement and experimentation

  • Interruption cost

    Interruption cost is a way to organize, support, or evaluate software work so that teams can make useful progress with less avoidable friction. It is most valuable when connected to a concrete outcome and the local conditions of the team using it.

    Developer productivity

  • Interruption rate

    Interruption rate is a developer productivity concept that helps teams understand interruption rate in the context of software delivery.

    Developer productivity

  • Inverse probability weighting

    Inverse probability weighting is a statistical or measurement concept used to describe, compare, or interpret engineering data. Its meaning depends on the unit of analysis, data-generating process, and question being asked.

    Measurement and experimentation

  • Item size

    Item size describes the work unit used in flow analysis. It may use changed behavior, files, stories, or another local convention, but comparisons require a consistent definition and awareness of complexity.

    Flow and capacity planning

  • Iteration planning

    Iteration planning is a concept used in software delivery planning to describe a condition, relationship, estimate, or decision about engineering work.

    Flow and capacity planning

  • Jailbreak

    Jailbreak is a language-model concept about representation and similarity. It names a mechanism, representation, prompting pattern, decoding control, or context behavior that can change how an AI system processes input and produces output.

    LLM fundamentals

  • Judge agreement

    Judge agreement is a defined lens for examining AI system behavior with a stated task, evidence, and interpretation rule.

    Evaluations and benchmarks

  • Judge bias

    Judge bias is a defined lens for examining AI system behavior with a stated task, evidence, and interpretation rule.

    Evaluations and benchmarks

  • Kanban board

    A Kanban board is a visualization of a workflow that makes work, states, WIP, and movement visible. It can be physical or digital, but the board is useful only when its columns reflect the real system.

    Flow and capacity planning

  • Kanban system

    A Kanban system manages knowledge work by making the workflow and constraints visible, limiting work in progress, pulling work according to capacity, and improving policies through feedback.

    Flow and capacity planning

  • Kernel fusion

    Kernel fusion is the serving concept concerned with kernel fusion during AI inference.

    Inference performance

  • Knowledge sharing

    Knowledge sharing is a way to organize, support, or evaluate software work so that teams can make useful progress with less avoidable friction. It is most valuable when connected to a concrete outcome and the local conditions of the team using it.

    Developer productivity

  • Knowledge transfer

    Knowledge transfer is a way to organize, support, or evaluate software work so that teams can make useful progress with less avoidable friction. It is most valuable when connected to a concrete outcome and the local conditions of the team using it.

    Developer productivity

  • Knowledge work

    Knowledge work is a developer productivity concept that helps teams understand knowledge work in the context of software delivery.

    Developer productivity

  • Krippendorff's alpha

    Krippendorff's alpha is a defined lens for examining AI system behavior with a stated task, evidence, and interpretation rule.

    Evaluations and benchmarks

  • Kruskal wallis test

    Kruskal wallis test is a statistical or measurement concept used to describe, compare, or interpret engineering data. Its meaning depends on the unit of analysis, data-generating process, and question being asked.

    Measurement and experimentation

  • Kurtosis

    Kurtosis is a statistical or measurement concept used to describe, compare, or interpret engineering data. Its meaning depends on the unit of analysis, data-generating process, and question being asked.

    Measurement and experimentation

  • KV cache

    KV cache is the serving concept concerned with kv cache during AI inference.

    Inference performance

  • Lack of cohesion

    A family of measures that estimates how weakly the responsibilities or data uses within a class are connected.

    Code quality and technical debt

  • Language-model objective

    Language-model objective is a language-model concept about mechanism and information flow. It names a mechanism, representation, training practice, or operational behavior that can change how an AI system processes input and produces output.

    LLM fundamentals

  • Late-arriving data

    Late-arriving data describes records delivered after their event window.

    Engineering analytics

  • Latency breakdown

    Latency breakdown is the serving concept concerned with latency breakdown during AI inference.

    Inference performance

  • Latency budget

    Latency budget is the serving concept concerned with latency budget during AI inference.

    Inference performance

  • Latency evaluation

    Latency evaluation is a defined lens for examining AI system behavior with a stated task, evidence, and interpretation rule.

    Evaluations and benchmarks

  • Latency-adjusted AI cost

    Latency-adjusted AI cost is a cost comparison that accounts for the operational impact of response time. It gives teams a way to name, measure, or reason about an economic property of an AI workload without treating raw usage as proof of value.

    Token costs and AI ROI

  • Latent representation

    Latent representation is a language-model concept about generation behavior and sampling. It names a mechanism, representation, prompting pattern, decoding control, or context behavior that can change how an AI system processes input and produces output.

    LLM fundamentals

  • Law of Demeter

    A guideline that limits how many unrelated object relationships a method navigates directly.

    Code quality and technical debt

  • Law of large numbers

    Law of large numbers is a statistical or measurement concept used to describe, compare, or interpret engineering data. Its meaning depends on the unit of analysis, data-generating process, and question being asked.

    Measurement and experimentation

  • Layer normalization

    Layer normalization is a language-model concept about evaluation design and failure analysis. It names a mechanism, representation, training practice, or operational behavior that can change how an AI system processes input and produces output.

    LLM fundamentals

  • Lead time

    Lead time is the elapsed time between a request entering a workflow and the requested outcome being delivered. In software teams, the start may be prioritization or commitment and the finish may be deployment, so the definition must be stated explicitly.

    Flow and capacity planning

  • Lead-time distribution

    A lead-time distribution shows how elapsed time varies from the chosen demand boundary to the chosen completion boundary. It describes the range and frequency of outcomes rather than one representative average.

    Flow and capacity planning

  • Leader election

    Leader election is a reliability concept used to describe a specific condition, control, or decision in the operation of software services.

    Reliability and observability

  • Learning rate schedule

    Learning rate schedule is a language-model concept about evaluation design and failure analysis. It names a mechanism, representation, training practice, or operational behavior that can change how an AI system processes input and produces output.

    LLM fundamentals