Converts Machine Learning Product Requirements Documents into structured JSON backlogs for iterative development and experimentation.
The Ralph skill bridges the gap between high-level machine learning requirements and actionable execution by converting markdown or text-based PRDs into a standardized prd.json format. Designed for the ML-Ralph framework, it enforces software engineering best practices within data science workflows, ensuring that tasks are sized for single iterations, dependencies are correctly sequenced, and every story includes verifiable acceptance criteria like Ruff, Mypy, and Pytest checks. This skill is essential for teams aiming to maintain rigorous logging and reproducibility in their ML pipelines.
주요 기능
01Enforcement of iterative story sizing
020 GitHub stars
03Pre-configured code quality acceptance criteria
04Automated conversion of text PRDs to JSON
05Dependency-aware task sequencing
06Structured evidence and metric logging templates
사용 사례
01Translating ML research goals into a structured development backlog
02Standardizing acceptance criteria across data science experiments
03Managing complex ML dependencies from scaffolding to operationalization