Large Language Models (LLMs) are rapidly reshaping the landscape of scientific research and industrial applications. The two-day training “LLMs in Science & Industry” offers a comprehensive and practical introduction to the core concepts, methods, and tools behind modern LLMs. It focusses on their effective and responsible use in scientific and industrial contexts.

The training will take place on May 12 and 13, 2025
at ScaDS.AI Dresden/Leipzig, Humboldstr. 25, 04105 Leipzig.

Registration

The participation fee includes all sessions and workshops as well as lunch, snacks and the social evening. Accommodation and travel can unfortunately not be covered by the registration fee and require self-organization.

Program

Participants will gain a solid foundation in deep learning and LLM architectures, as well as explore techniques for efficient training and inference. Furthermore they will learn how to fine-tune and evaluate models for domain-specific tasks. The course also addresses legal and ethical considerations, data handling, and cutting-edge approaches such as Retrieval-Augmented Generation (RAG).

The program covers the following topics:

• Deep Learning & Large Language Models Basics
• Training Large Language Models, Scaling laws
• Resource efficient inference, Distillation, Pruning
• Fine-tuning an LLM
• Large Language Model Evaluation
• Data Sets & legal aspects of LLMs
• Retrieval Augmented Generation (RAG)
• Work on you own data Sessions

Schedule

A schedule will be published here soon.

Methodological Approach

The two-day program combines expert-led lectures with hands-on sessions, including the opportunity to work with your own data. A dedicated networking event will provide space for exchange and collaboration across disciplines.

Target Audience

This training is designed for researchers, engineers and professionals.

Confirmed Speakers

• Prof. Dr. Michael Färber (ScaDS.AI Dresden/Leipzig, TU Dresden)
• Prof. Dr. Simon Razniewski (ScaDS.AI Dresden/Leipzig, TU Dresden)
• Dr. Aaron Klein (ScaDS.AI Dresden/Leipzig, Leipzig University)
• Dr. Robert Haase (ScaDS.AI Dresden/Leipzig, Leipzig University)

Organizers

Dr. Robert Haase (ScaDS.AI Dresden/Leipzig, Leipzig University)