Course
Digicomp Code RAG
Retrieval-Augmented Generation (RAG) Chatbots – Implementation and Deep Dive («RAG»)
Course facts
- Practical experience in implementing RAG chatbots
- Architecture of RAG systems
- Evaluating and optimizing RAG chatbots
- Security of RAG chatbots
- Information retrieval for text documents, including text search and semantic search using vector databases
In this course, participants will acquire both the necessary theoretical knowledge and practical experience to implement production-ready RAG chatbots in an enterprise environment. The course content includes:
1 Implementation of a production-ready RAG chatbot using selected open source tools
Together, we will implement a production-ready RAG chatbot based on Python-based open source libraries and selected public cloud services for hosting language models.
2 Architecture and functionality of Retrieval Augmented Generation (RAG) chatbots
Functionality and architectures for RAG chatbots; scaling and cost aspects.
3 Fundamentals of information retrieval and semantic search for text documents
Classic text search and semantic search using vector databases; how vector search engines work.
4 Evaluation and optimization of RAG systems
Systematic evaluation and optimization of RAG chatbots.
5 Security of RAG systems
Developing a holistic security concept for RAG; the role of red teaming, content filtering, prompt logging, etc.
6 Extensions of RAG systems
Overview of further developments and extensions of standard RAG chatbots such as Agentic AI and Agentic RAG, multimodal RAG, or text-to-query.
- Hands-on implementation
- Interactive theory units
- Exchange of experiences within the group
- Software engineers
- Data scientists, AI developers, ML engineers
- IT architects (e.g., software architects, enterprise architects)
- Product owners, product managers in the AI field with experience in software development
Basic knowledge of software development and Python programming is required. Experience with AI technologies or cloud platforms is an advantage, but not essential.