Automate design, deployment, and maintenance of production ML systems.
npx skills add https://github.com/dinoudon/sdlc-skills --skill ml-engineering-dinoudon
Please help me install this Agent Skill. Skill: ml-engineering Source: https://github.com/dinoudon/sdlc-skills/tree/main/full/ml-engineering Command: npx skills add https://github.com/dinoudon/sdlc-skills --skill ml-engineering-dinoudon
ML systems often struggle with bridging data, experiments, deployment, and monitoring. This Skill provides a mature, end-to-end framework for MLOps, covering data versioning, feature stores, model serving, and drift monitoring to keep models reliable in production.
Outline an end-to-end ML deployment plan for a production model, including data lineage, feature store design, model serving, monitoring, and retraining triggers.
High-intent search queries and answers about installing and using this skill.
To build an end-to-end ML pipeline, you must design data lineage, configure a feature store, set up model serving, establish drift monitoring, and define automated retraining triggers to maintain production reliability.
Drift detection monitors production data for statistical changes that degrade model performance. You need it for machine learning systems when deployed models face evolving data distributions, requiring automated retraining triggers to maintain accuracy.
Setting up a feature store involves designing centralized storage for curated features that serve both training and inference. This ensures data consistency across ML pipelines and bridges data versioning with model serving workflows.
Yes, the MLOps framework supports experiment tracking and model registry workflows. It provides maturity modeling and tooling guidance to manage scalable ML projects, ensuring reliable deployment and monitoring guardrails for production models.
The best way to monitor machine learning models for drift is implementing automated guardrails that track data and prediction distributions. This enables proactive detection of degradation, triggering automated retraining pipelines to restore reliability.
Machine learning systems struggle with deployment and monitoring because bridging data, experiments, and production environments is complex. Without a mature MLOps framework covering feature stores and drift monitoring, models become unreliable over time.
ancolemanBuild RAG pipelines, feature stores, and embedding generation for AI/ML systems.
nahisahoDetect data and concept drift in production ML models.
belokonmOrchestrate end-to-end MLOps workflows for building, deploying, and monitoring production ML systems.
augustoheissProvide end-to-end machine learning engineering support for deploying production models.
