Sebastian Simon - Portfolio

About Me

Hey! I'm Sebastian, a PhD student and research assistant at Leipzig University with a strong interest in the configuration landscape of modern software systems and configuration dependencies.

A key part of my research focuses on modeling the configuration landscape of modern software systems to uncover cross-technology configuration dependencies. I also work with large language models (LLMs), retrieval-augmented generation (RAG), and agent-based systems to build tools for detecting and validating configuration dependencies.

Outside of research, I enjoy staying active—bouldering, playing badminton, and going for runs or to the gym are some of my favorite ways to unwind.

⚡ SUPERPOWERS ⚡

9001

Technology Stack

Proficient in a range of programming languages and technologies including Python, Java, C#, Node.js, Spring, Maven, and Docker.

8500

Artificial Intelligence

Hands-on experience with LLMs, embedding models, and advanced architectures such as RAG and agent-based systems.

9200

Research

Extensive experience in empirical research, academic publishing, presenting, and applying research insights to practical software prototypes.

8800

Project Organization

Skilled in supervising students and organizing courses such as software engineering internships and and SE4AI.

Professional Journey

Education/Work

2020 - Present

PhD Student/Research Assistant

Leipzig University

Modelling configuration landscapes of modern software systems and tackling configuration dependencies.

Software Engineering AI/ML Python

Education

2018 - 2022

M.Sc. Computer Systems in Engineering

Otto-von-Guericke University Magdeburg

Thesis on finding test gaps in software systems using changed-based analysis.

Software Engineering Python

Work

Apr 2019 - Sep 2020

Student Assistant

Selectline Software

Web application testing with Cypress.io and C# application development.

Cypress JavaScript C#

Work

Apr 2018 - Mar 2019

Student Assistant

SensoTech GmbH

Web application testing and automation using Selenium.

Selenium C#

Internship

Sep 2017 - Dec 2017

Intern

Volkswagen AG

Web interface testing and test automation with HP UFT.

HP UFT Visual Basic

Education

2014 - 2018

B.Sc. Computer Systems in Engineering

Otto-von-Guericke University Magdeburg

Thesis on automated cross-browser testing

Java Python Algorithms

Featured Projects

CfgNet

A framework for modeling the configuration landscape of software projects to detect and extract cross-stack configuration dependencies.

Python Poetry

GitHub

RepoMiner

As part of my master thesis, I developed a tool to analyze the historical changes in software repositories to identify test gaps.

Python Flask PyDriller

GitHub

DiffAgent

A multi-agent system that validates configuration files changes to prevent misconfigurations before they reach production (WIP).

Python Langchain Langgraph OpenAI

GitHub

LastSoul

A 2D survival game where a lost soul needs to find his way back to its original body (WIP).

C# Unity

GitHub

Research & Publications

On Automating Configuration Dependency Validation via Retrieval-Augmented Generation

S. Simon, A. Mailach, J. Dorn, N. Siegmund

IEEE/ACM 40th International Conference on Automated Software Engineering, ASE 2025

We validated configuration dependencies via RAG, showing that incorporating tailored contextual information significantly improves the validation performance of all studied LLMs.

PDF

Themes of Building LLM-Based Applications for Production: A Practitioner's View

A. Mailach, S. Simon, J. Dorn, N. Siegmund

IEEE/ACM 4th International Conference on AI Engineering–Software Engineering for AI (CAIN, 2025)

We mapped topics that practitioners discuss online about building LLM-based applications, offering practical insights into key considerations for developing such systems.

PDF

A Methodology for Evaluating RAG Systems: A Case Study On Configuration Dependency Validation

S. Simon, A. Mailach, J. Dorn, N. Siegmund

ArXiv

This paper presents a methodology for a sound and reliable evaluation of RAG systems. We demonstrated its applicability on a real-world software engineering research task: the validation of configuration dependencies.

PDF Code

Exploring Hyperparameter Usage and Tuning in Machine Learning Research

S. Simon, N. Kolyada, C. Akiki, M. Potthast, B. Stein, N.Siegmund

IEEE/ACM 2nd International Conference on AI Engineering–Software Engineering for AI (CAIN, 2023)

This work explores hyperparameter usage and tuning in machine learning research. We showed that the most of the hyperparameters remain untouched, and those that have been changed use constant values. We also found a significant difference between tuning hyperparameters and the reporting about it in the corresponding research papers.

PDF Code

CfgNet: A Framework for Tracking Equality-Based Configuration Dependencies Across a Software Project

S. Simon, N. Ruckel, N. Siegmund

IEEE Transactions on Software Engineering

We developed CfgNet, a framework that models the configuration landscape of a software project as a configuration network in an extensible and artifact-independent way. With CfgNet, we enable the early detection of possible dependency violations and proactively prevent misconfigurations during software development and maintenance.

PDF Code