Auto-sync: 2026-04-21 00:02

wiki/concepts/ML-Ops.md

@@ -1,23 +1,25 @@
 ---
 title: "ML Ops"
 type: concept
-tags: [DevOps, ML, operations]
-sources: [public-cloud-learning-sessions-introduction-to-artificial-intelligence-ai-machine-learning-20240206]
-last_updated: 2024-02-06
+tags: [machine-learning, operations, lifecycle]
+sources: [specialized-model-qa]
+last_updated: 2026-04-20
 ---
 
-## Summary
-ML Ops 结合机器学习和运营，涉及人员、技术和流程，以实现协作式 ML 解决方案。
-
 ## Definition
-ML Ops (Machine Learning Operations) 是将 DevOps 原则应用于机器学习系统的实践，包括数据管道、训练管道和推理管道的自动化和监控。
+ML Ops is the discipline of operationalizing machine learning models across development, deployment, monitoring, and governance.
 
-## Key Attributes
-- **核心组成**：数据管道、训练管道、推理管道
-- **关注点**：数据溯源、模型管理、部署工作流
-- **工具**：Amazon SageMaker、MLflow、Kubeflow
+## Core Areas
+- Data pipelines
+- Training and deployment
+- Monitoring and drift detection
+- Governance and auditability
 
-## Connections
-- [[DevOps]] ← extends ← [[ML Ops]]
-- [[ML Ops]] ← manages ← [[Machine Learning]]
-- [[Amazon SageMaker]] ← implements ← [[ML Ops]]
+## Relevance to Model QA
+- Provides the operational context for audits
+- Supplies monitoring and reproducibility artifacts
+- Supports remediation and retraining loops
+
+## Related Concepts
+- [[Model Audit]]
+- [[Discrimination Metrics]]