Towards an automated teaching assistant for network configuration

Authors: Pietro Ronchetti*, Jean Mégret*, and Laurent Vanbever
Report
*equal contribution
PDF

Abstract

The adoption of LLM tools by students for edu cational assignments has risen in popularity. However, general purpose tools hit their limits when the topic is too niche and require contextualized information about the lecture or risk not aligning with the teaching intents of the course. In an effort to improve these limitations, ubiquitous in higher degree education, we propose to augment the capabilities of generic LLMs by integrating lecture specific content, tooling and pedagogical guidelines. We present a preliminary implementation of these features and investigate its effectiveness in a case study.

People

Pietro Ronchetti
PhD student
Jean Mégret
PhD student

BibTex

@misc{ronchetti2025towards,
  author    = {Ronchetti, Pietro and M{\'{e}}gret, Jean Bernard David and Vanbever, Laurent},
  title     = {{Towards an automated teaching assistant for network configuration}},
  year      = 2025,
  month     = nov,
  publisher = {ETH Zurich},
  doi       = {10.3929/ETHZ-C-000806383},
  url       = {https://www.research-collection.ethz.ch/bitstream/handle/20.500.11850/806383/TABot.pdf}
}

Research Collection: 20.500.11850/806383