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 termsLeast-loaded routing
Least-loaded routing is a model-routing or gateway concept used to manage traffic selection for AI requests. It describes a distinct decision, control, interface, or observation point between an application and one or more model providers.
Model routing and gatewaysLegacy modernization
Legacy modernization 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 debtLegacy system
Legacy system 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 debtLength penalty
Length penalty 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 fundamentalsLines of code
Lines of code is a count of source lines in a file, change, or codebase. Depending on the tool, the count may include blank lines, comments, generated files, or only executable statements, so the definition must be stated.
Developer productivityLiskov substitution principle
A design principle that requires a subtype to remain valid wherever its declared base type is expected.
Code quality and technical debtListwise deletion
Listwise deletion 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 experimentationListwise ranking
Listwise ranking is a defined lens for examining AI system behavior with a stated task, evidence, and interpretation rule.
Evaluations and benchmarksLiveness probe
Liveness probe is a test of whether a process should be restarted because it no longer functions at a basic level.
Reliability and observabilityLLM inference
LLM inference is the process of running a trained language model on an input to produce an output. For a text-generating model, it typically involves processing the input context and generating additional tokens according to a decoding strategy.
LLM fundamentalsLLM latency
LLM latency is the time associated with receiving a language model response. Common measures include time to first token, time between generated tokens, and time to the final token, each describing a different user experience.
Inference performanceLoad shedding
Load shedding is a model-routing or gateway concept used to manage reliable request governance for AI requests. It describes a distinct decision, control, interface, or observation point between an application and one or more model providers.
Model routing and gatewaysLoad testing
Load 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 debtLocalization testing
Localization 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 debtLog enrichment
Log enrichment is adding context to a log record such as service identity, deployment, or correlation fields.
Reliability and observabilityLog level
Log level is a classification such as debug, info, warning, or error indicating intended operational importance.
Reliability and observabilityLog normal distribution
Log normal distribution 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 experimentationLog parsing
Log parsing is conversion of raw log text or fields into a structured searchable representation.
Reliability and observabilityLog retention
Log retention is the policy determining how long logs remain available and under what archive or deletion rules.
Reliability and observabilityLog signal
Log signal is a timestamped record of an event or state transition emitted by software or infrastructure.
Reliability and observabilityLog-based alert
Log-based alert is a trigger based on matching log patterns, counts, rates, or structured conditions.
Reliability and observabilityLogit lens
Logit lens 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 fundamentalsLong context window
Long context window is the serving concept concerned with long context window during AI inference.
Inference performanceLong-context evaluation
Long-context evaluation is a defined lens for examining AI system behavior with a stated task, evidence, and interpretation rule.
Evaluations and benchmarksLongitudinal study
Longitudinal study 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 experimentationMain branch health
Main branch health is a software delivery concept used to describe a specific event, interval, control, or operating condition in the path from source change to production behavior. A useful definition names the boundary, unit, and decision the measure supports.
DORA and DevOpsMaintainability risk
The likelihood that future changes to a software area will be costly, error-prone, or difficult to validate.
Code quality and technical debtMaintainability testing
Maintainability 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 debtMaintenance work
Maintenance work is a developer productivity concept that helps teams understand maintenance work in the context of software delivery.
Developer productivityMaker schedule
Maker schedule 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 productivityMann-Whitney U test
Mann-Whitney U 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 experimentationManual instrumentation
Manual instrumentation records domain-specific events explicitly.
Engineering analyticsMargin of error
Margin of error 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 experimentationMarginal AI cost
Marginal AI cost is the additional expense caused by one more request, token, user, or workflow. 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 ROIMasked language modeling
Masked language modeling is a language-model concept about serving behavior and operational tradeoffs. It names a mechanism, representation, training practice, or operational behavior that can change how an AI system processes input and produces output.
LLM fundamentalsMasked span corruption
Masked span corruption is a language-model concept about training behavior and measurement. It names a mechanism, representation, training practice, or operational behavior that can change how an AI system processes input and produces output.
LLM fundamentalsMaturity benchmark
Maturity benchmark is an analytical concept for using reference values to understand engineering performance, variation, or capability.
Engineering analyticsMax concurrent requests
Max concurrent requests is the serving concept concerned with max concurrent requests during AI inference.
Inference performanceMax pooling
Max pooling 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 fundamentalsMax tokens
Max tokens is the serving concept concerned with max tokens during AI inference.
Inference performanceMaximum generation length
Maximum generation length 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 fundamentalsMean cycle time
Mean cycle time is calculated by adding the cycle times in a selected population and dividing by the number of items. It can summarize a dataset, but a few long-running items may pull it away from what most work experiences.
Flow and capacity planningMean pooling
Mean pooling 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 fundamentalsMean time to mitigate
Mean time to mitigate is a software delivery concept used to describe a specific event, interval, control, or operating condition in the path from source change to production behavior. A useful definition names the boundary, unit, and decision the measure supports.
DORA and DevOpsMean time to resolve
Mean time to resolve is a software delivery concept used to describe a specific event, interval, control, or operating condition in the path from source change to production behavior. A useful definition names the boundary, unit, and decision the measure supports.
DORA and DevOpsMean time to restore
Mean time to restore is the average elapsed time needed to restore a service after an incident or failure. The result depends on which events count, how recovery is defined, and whether the distribution is summarized by a mean or another statistic.
DORA and DevOpsMeasurement bias
Measurement bias describes systematic distortion in what a metric represents.
Engineering analyticsMeasurement context
Measurement context what context makes an engineering measure interpretable.
Engineering analyticsMeasurement error
Measurement error 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 experimentationMeasurement framework
Measurement framework governs how engineering measures are chosen and interpreted.
Engineering analytics