# RoboArchitect: Automating the Design of Robotic Agents with Persistent Memory

- Source: HuggingFace Daily Papers
- Published: 2026-10-30T00:00:00.000Z
- RoboHot: https://robohot.news/en/items/aq8052xdzj0qfaec67hrtw15y
- Original: https://huggingface.co/papers/2610.42141

## Summary

RoboArchitect lets a coding model observe robot rollouts, diagnose failures, and iteratively revise the agent harness around a frozen policy. Without updating any model weights, it raises average success from 9.7% to 43.3% (4.5x) on unseen scenes and from 12.7% to 40.2% (3.2x) on unseen tasks.

## Why it matters

It automatically evolves control code and persistent memory outside a frozen policy, offering a transferable approach to improving manipulation success without finetuning weights.