ralph-experiment

What problem does it solve?

Hypothesis-driven experiment loop that structures iteration, tracking, and learning using a notebook and Git-based state machine.

Core Features & Use Cases

• Structured experimentation with hypotheses, predictions, and learnings
• Git-backed checkpoints and reverts to keep experiments isolated
• Agent delegation to distribute work across executors, analysts, and architects
• Cross-iteration notebook for capturing insights and guiding next steps

Quick Start

Define an initial hypothesis, run a baseline measurement, and begin iterating with one hypothesis at a time, logging results in the local experiment notebook.