Language Models in Practice: Using LLMs Safely, Efficiently, and Locally
General Information
This microcredential provides a comprehensive and practice-oriented introduction to Large Language Models (LLMs). Participants will gain foundational knowledge of LLM architectures, training methods, and application scenarios - from general use cases to secure, local deployments. Through hands-on exercises they will learn effective prompting techniques, work with open-source frameworks, and explore methods to enhance model efficiency and ensure safe operation. The course bridges theory and real-world application, preparing students and professionals to use LLMs responsibly and effectively in both academic and business contexts.
Nevertheless, good Python programming skills are required. Basic knowledge on deep neural networks is advantageous. The course language is English.
Registration and materials
https://bildungsportal.sachsen.de/opal/auth/RepositoryEntry/50362089486
Content
- Inside the Black Box: How Transformers and Embeddings Work
- One Model, Many Skills: Exploring LLM Types and Applications
- Shaping Intelligence: Datasets, Benchmarks, and Fine-Tuning Methods
- Smarter Interactions: From Prompts to Knowledge-Enhanced Generation
- Hands-on with LLMs: Experimenting with Prompts and Retrieval
- Trustworthy AI: Risks, Attacks, and Defenses
- LLM Efficiency: Optimizing Speed, Size, and Resources
- From Models to Agents: Planning, Tools, and Action
Course plan
Alternating online courses and offline self-study
