MLOps 工程师

# MLOps Engineer (MLOps 工程师) ## Overview **MLOps Engineer (MLOps 工程师)** is a senior MLOps specialist for building and maintaining ML platforms. It covers infrastructure automation, CI/CD pipelines, model versioning, experiment tracking, resource orchestration, monitoring, security, and cost optimization — creating scalable, reliable ML infrastructure that accelerates model delivery. ## Key Capabilities - ✅ **Platform architecture** — infrastructure design, components, monitoring design - ✅ **CI/CD for ML** — pipeline automation, validation, artifact management, rollbacks - ✅ **Model versioning** — registry, lineage, reproducibility, rollback capability - ✅ **Experiment tracking** — parameter/metric logging, comparison, collaboration - ✅ **Resource orchestration** — Kubernetes, GPU scheduling, auto-scaling, cost optimization - ✅ **Automation & reliability** — IaC, GitOps, blue-green/canary, chaos engineering, multi-cloud ## Usage 1. Load the `mlops-engineer` agent into your Markus workspace. 2. @mention `@mlops-engineer` (or route via workflow) with platform requirements and team needs. 3. It assesses the current state, builds automation and pipelines, configures monitoring and security, and returns a documented, cost-optimized ML platform. ## Licensing Source Based on the **mlops-engineer** subagent from [awesome-claude-code-subagents](https://github.com/VoltAgent/awesome-claude-code-subagents) (MIT). --- **Attribution:** Based on [mlops-engineer subagent](https://github.com/VoltAgent/awesome-claude-code-subagents/blob/main/categories/05-data-ai/mlops-engineer.md) from [awesome-claude-code-subagents](https://github.com/VoltAgent/awesome-claude-code-subagents) - Original author: VoltAgent - License: MIT