Categories
Pages
-

DBIS

Using LLMs to Support Modernization of Legacy Systems

August 3rd, 2023

In recent years, the advancements in natural language processing and machine learning have revolutionized various industries and domains.

This thesis aims to explore the potential of training a Large Language Model (LLM) on Git source code repositories to enable it to effectively respond to queries regarding the codebase, such as dependencies and functionality.

Thesis Type
  • Bachelor
Student
Ilija Kovacevic
Status
Finished
Presentation room
Seminar room I5 6202
Supervisor(s)
Stefan Decker
Advisor(s)
Michal Slupczynski
Contact
slupczynski@dbis.rwth-aachen.de

The main goal of this thesis is to develop and fine-tune a Large Language Model capable of comprehending and answering questions related to Git-based source code repositories.

The proposed LLM-based approach holds promise for enhancing documentation, facilitating modernization efforts of legacy systems, and fostering better stakeholder involvement in machine learning projects.

This research aims to contribute to the software engineering and machine learning fields in multiple ways. The developed LLM has the potential to serve as an intelligent tool for developers, project managers, and stakeholders in understanding and interacting with code repositories.

Co-Advisor: Ada Slupczynski (SWC RWTH)


Prerequisites:
  • DevOps/MLOps
  • Python