Retrieval-Augmented Generation (RAG) Training in Kraków, Poland

The Retrieval-Augmented Generation (RAG) Training is a hands-on program focused on building accurate, scalable, and enterprise-ready GenAI applications by combining Large Language Models (LLMs) with external knowledge sources.
Participants learn how to design RAG pipelines that reduce hallucinations, improve answer quality, and leverage private data securely using modern AI stacks.

Introduction to RAG & Enterprise GenAI

• Why RAG is needed for enterprise AI

• Limitations of standalone LLMs

• RAG vs fine-tuning

RAG Architecture & Workflow

• High-level RAG system design

• Data ingestion → retrieval → generation flow

• Online vs offline RAG pipelines

• Stateless vs stateful RAG systems

Document Ingestion & Data Preparation

• Handling PDFs, Word, HTML, and text files

• Cleaning and preprocessing text

• Metadata extraction and enrichment

• Versioning and data updates

Chunking Strategies & Indexing

• Fixed-size vs semantic chunking

• Overlapping chunks and context windows

• Chunk size impact on accuracy and cost

• Indexing best practices

Embeddings & Vector Databases

• What are embeddings and how they work

• Creating embeddings at scale

• Vector indexing and similarity search

• Performance and cost considerations

Retrieval Strategies

• Similarity search vs hybrid search

• Keyword + vector retrieval

• Filtering using metadata

• Ranking and re-ranking techniques

Prompt Engineering for RAG

• Prompt templates for grounded responses

• Injecting retrieved context safely

• Reducing hallucinations

• System prompts vs user prompts

RAG with Frameworks

• Building RAG pipelines using LangChain

• Using LlamaIndex for document indexing

• Comparing frameworks and trade-offs

• Modular and extensible pipeline design

Advanced RAG Techniques

• Multi-document and multi-hop RAG

• Conversational RAG with memory

• Agentic RAG workflows

• Query rewriting and expansion

Evaluation & Optimization

• Measuring RAG quality and relevance

• Latency, cost, and accuracy trade-offs

• Human-in-the-loop evaluation

• Continuous improvement strategies

Security, Governance & Responsible AI

• Protecting sensitive enterprise data

• Access control and authorization

• Data leakage prevention

• Compliance and audit considerations

Deployment & Scaling RAG Systems

• API-based RAG services

• Caching and performance optimization

• Scaling retrieval and inference

• Monitoring and observability

Hands-on Exercises

Summary and Conclusion