Git+AI Training

This page contains training materials for using Git in combination with modern AI-Agents such as

• Github Copilot

• Openhands / All-Hands

• Git-Bob (developed in Leipzig, available on the Scientific Computing Gitlab sovereign IT infrastructure of the University Compute Center URZ of Leipzig University)

The main rationale of using the AI-agents in the browser is to document reproducibly which parts of a work has been done by humans and which part was done by an AI-agent. Another advantage of this approach is that we do not need to download anything, we typically can try out things in the browser first.

Target audience and teaching goal

The exercises are written for scientists with basic experience in programming and with git. We will explore AI-agent systems that can solve coding tasks and assist with scientific data analysis.

Attendees can choose an AI-Agent system and follow exercises step-by-step. Afterwards, goal is to explore the limits of the technology. Only if we know where it succeeds and fails we can use it responsibly.

Read more

Recommended literature:

• Jimenez et al “SWE-bench: Can Language Models Resolve Real-World GitHub Issues?”, ICLR 2024

• Wang et al. “OpenHands: An Open Platform for AI Software Developers as Generalist Agents”, ICLR 2025

• Haase “Towards transparency and knowledge exchange in AI-assisted data analysis code generation”, Nat. Comput. Sci 2025

Trainers

• Dr. Ilona Lipp, Leipzig University

• Dr. Robert Haase, ScaDS.AI Dresden/Leipzig

Acknowledgements

We acknowledge the financial support by the Federal Ministry of Education and Research of Germany and by Sächsische Staatsministerium für Wissenschaft, Kultur und Tourismus in the programme Center of Excellence for AI-research „Center for Scalable Data Analytics and Artificial Intelligence Dresden/Leipzig“, project identification number: ScaDS.AI