Saved in:
Bibliographic Details
Main Authors: Yang, Zonglin, Xie, J. -H., Zhang, Lining, Jia, Jiyou, Chen, Zhi-X.
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.20650
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908903472627712
author Yang, Zonglin
Xie, J. -H.
Zhang, Lining
Jia, Jiyou
Chen, Zhi-X.
author_facet Yang, Zonglin
Xie, J. -H.
Zhang, Lining
Jia, Jiyou
Chen, Zhi-X.
contents Deploying high-fidelity AI tutors in schools is often blocked by the Resource Curse -- the need for expensive cloud GPUs and massive data engineering. In this practitioner report, we present a replicable Standard Operating Procedure that breaks this barrier. Using a Vision-Language Model data cleaning strategy and a novel Shadow-RAG architecture, we localized a graduate-level Applied Mathematics tutor using only 3 person-days of non-expert labor and open-weights 32B models deployable on a single consumer-grade GPU. Our pilot study on a full graduate-level final exam reveals a striking emergence phenomenon: while both zero-shot baselines and standard retrieval stagnate around 50-60% accuracy across model generations, the Shadow Agent, which provides structured reasoning guidance, triggers a massive capability surge in newer 32B models, boosting performance from 74% (Naive RAG) to mastery level (90%). In contrast, older models see only modest gains (~10%). This suggests that such guidance is the key to unlocking the latent power of modern small language models. This work offers a cost-effective, scientifically grounded blueprint for ubiquitous AI education.
format Preprint
id arxiv_https___arxiv_org_abs_2603_20650
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle From 50% to Mastery in 3 Days: A Low-Resource SOP for Localizing Graduate-Level AI Tutors via Shadow-RAG
Yang, Zonglin
Xie, J. -H.
Zhang, Lining
Jia, Jiyou
Chen, Zhi-X.
Artificial Intelligence
Computers and Society
I.2.7; K.3.1
Deploying high-fidelity AI tutors in schools is often blocked by the Resource Curse -- the need for expensive cloud GPUs and massive data engineering. In this practitioner report, we present a replicable Standard Operating Procedure that breaks this barrier. Using a Vision-Language Model data cleaning strategy and a novel Shadow-RAG architecture, we localized a graduate-level Applied Mathematics tutor using only 3 person-days of non-expert labor and open-weights 32B models deployable on a single consumer-grade GPU. Our pilot study on a full graduate-level final exam reveals a striking emergence phenomenon: while both zero-shot baselines and standard retrieval stagnate around 50-60% accuracy across model generations, the Shadow Agent, which provides structured reasoning guidance, triggers a massive capability surge in newer 32B models, boosting performance from 74% (Naive RAG) to mastery level (90%). In contrast, older models see only modest gains (~10%). This suggests that such guidance is the key to unlocking the latent power of modern small language models. This work offers a cost-effective, scientifically grounded blueprint for ubiquitous AI education.
title From 50% to Mastery in 3 Days: A Low-Resource SOP for Localizing Graduate-Level AI Tutors via Shadow-RAG
topic Artificial Intelligence
Computers and Society
I.2.7; K.3.1
url https://arxiv.org/abs/2603.20650