Course Outline
Module I: Foundations of Large Language Models
1. How LLMs work – tokenization, context window, and the issue of hallucinations
2. Prompt engineering techniques for optimal results
3. Interacting with models via the OpenAI API
4. The LangChain framework in practice
5. Retrieval Augmented Generation (RAG) system architecture
Module II: Preparing Textual Documents
1. Extracting content from various file formats (PDF, DOCX, TXT)
2. The concept of text chunking
3. Document segmentation strategies
4. The impact of chunk size on system quality
Module III: Vectorization and Data Storage
1. Principles of vector representation of text (embeddings)
2. Semantic search based on vector similarity
3. The Qdrant vector database – configuration and application in RAG
4. Indexing and managing vector collections
Module IV: Document Retrieval and Quality Assessment
1. The retrieval process in RAG systems
2. Building context for LLM queries
3. Reranking techniques for results
4. The DeepEval framework for assessing RAG systems
Module V: Web Application for RAG Systems
1. Fundamentals of the Streamlit library
2. Implementing the user interface
3. Integrating RAG system components into a web application
