Artificial Inteligence

The course conceptualizes artificial intelligence as a system interconnecting models, data, decision-making mechanisms, and operational infrastructure. Emphasis is placed on reliability and explainability, reinforcement learning from basics to multi-agent approaches, system aspects, and a comparison of generative methods including evaluation and synthetic data generation. Additional focus is on decision-making frameworks for LLMs vs. SLMs (quality–latency–privacy), working with custom knowledge bases (RAG vs. fine-tuning), agent orchestration and the Model Context Protocol, as well as federated and privacy-preserving learning. The course further covers key areas such as production prompting, in-context learning and inference engineering, neuro-symbolic integration with knowledge graphs, and AutoML/meta-learning with emphasis on reproducibility. The course is practically oriented toward system integration, security, evaluation, and deployment of language models and multimodal pipelines. It builds on the courses Natural Language Processing (providing foundations of Transformer, BERT/GPT, and basic prompting) and Deep Learning (covering CNN/RNN/GAN architectures in depth).