How does this integrate with the SpecWeave framework?

It follows the SpecWeave lifecycle (spec-plan-tasks-implement-validate), linking every model iteration to a specific development increment and automatically updating architecture docs.

Can I use it for model explainability?

Absolutely. The skill includes dedicated stages for computing SHAP values or LIME explanations and documenting example predictions with rationales.

How does it help with model reproducibility?

It creates a dedicated increment folder for every ML task, automatically logging hyperparameters, metrics, artifacts, and environment requirements to ensure every run can be recreated.

Does this skill work with any ML framework?

Yes, it supports major frameworks including Scikit-Learn, PyTorch, TensorFlow/Keras, and XGBoost/LightGBM through specific integration points.

Does it handle data preprocessing?

Yes, it includes specialized stages for data exploration (EDA), quality assessment, schema validation, and feature engineering pipelines.

ML Pipeline Orchestrator

Name: ML Pipeline Orchestrator
Author: anton-abyzov

byanton-abyzov

•

Ciencia de Datos y ML

Orchestrates end-to-end machine learning workflows within a disciplined, spec-driven development framework.

Acerca de

This skill transforms machine learning development into a structured workflow by integrating the SpecWeave autonomous development framework. It guides users through the entire ML lifecycle—from data exploration and feature engineering to model training and deployment—ensuring every experiment is tracked, documented, and reproducible. By treating ML features as managed increments, it bridges the gap between experimental data science and production-grade software engineering, automatically generating living documentation and maintaining model version traceability.

Características Principales

Built-in model evaluation and explainability artifacts (SHAP/LIME)
13 GitHub stars
Seamless synchronization with living documentation and architecture records
Multi-framework support for Scikit-Learn, PyTorch, TensorFlow, and XGBoost
Integrated experiment tracking with automated metric logging and reporting
Automated ML increment structure including specs, plans, and task lists

Casos de Uso

Standardizing ML workflows across a team using the SpecWeave framework
Building production-ready recommendation or fraud detection systems
Ensuring complete reproducibility and versioning for experimental models

Acerca de

Características Principales

Built-in model evaluation and explainability artifacts (SHAP/LIME)
13 GitHub stars
Seamless synchronization with living documentation and architecture records
Multi-framework support for Scikit-Learn, PyTorch, TensorFlow, and XGBoost
Integrated experiment tracking with automated metric logging and reporting
Automated ML increment structure including specs, plans, and task lists

Casos de Uso

Standardizing ML workflows across a team using the SpecWeave framework
Building production-ready recommendation or fraud detection systems
Ensuring complete reproducibility and versioning for experimental models