DeepYard
R

ReASearch

Framework for autonomous reasoning-driven optimization across prompts, programs, and ML workflows

Open SourceFree

About

ReASearch is an open-source framework that enables AI agents to autonomously optimize their own reasoning processes by searching across prompts, programs, and ML workflows. Unlike traditional approaches that rely on external controllers, it internalizes the search policy directly into the agent, allowing for self-directed improvement. Designed for researchers exploring meta-learning and autonomous agent optimization.

Details

Language
Patterns

Tags

autonomousframeworkopen-sourceoptimizationresearch