II.
Workflow overview
Reference · liveworkflow:llm-eval-pipeline
LLM Evaluation Pipeline overview
Operates the continuous evaluation pipeline for LLM-powered features — maintaining eval datasets, running benchmark suites across model versions, tracking quality trends, and producing comparative reports. Excludes model fine-tuning.
Attributes
displayName
LLM Evaluation Pipeline
workflowKind
data
triggerType
scheduled
typicalCadence
weekly
complexity
cross-team
description
Operates the continuous evaluation pipeline for LLM-powered features —
maintaining eval datasets, running benchmark suites across model versions,
tracking quality trends, and producing comparative reports. Excludes model fine-tuning.
Outgoing edges
applies_to_domain2
- domain:ml-ops·DomainMLOps
- domain:data-science·DomainData Science
involves_role3
- role:ml-engineer·RoleMachine Learning Engineer
- role:ai-champion·RoleAI Champion
- role:data-scientist·RoleData Scientist
performed_by_org_unit3
- org-unit:ml-team·OrgUnitML Team
- org-unit:ai-enablement·OrgUnitAI Enablement
- org-unit:research-engineering·OrgUnitResearch Engineering
requires_skill_area3
- skill-area:eval-driven-development·SkillAreaEval-Driven LLM Development
- skill-area:prompt-engineering·SkillAreaPrompt Engineering
- skill-area:python-data-pipelines·SkillAreaPython Data Pipelines
triggers_responsibility2
- responsibility:ai-tooling-evaluation·ResponsibilityAI Tooling Evaluation
- responsibility:data-quality-monitoring·ResponsibilityData quality monitoring
Incoming edges
None.