Crash Course: Building a RAG system

Overview

This crash course offers a practical, hands-on introduction to building Retrieval-Augmented Generation (RAG) systems using real data sources. The program focuses on assembling a complete RAG workflow that operates locally, without exposing private documents to external networks. Participants will learn how to combine semantic search, embeddings, vector databases, and large language models to answer questions based on their own documents, with sources cited. The course emphasizes practical application for data scientists, analysts, AI enthusiasts, and technical leaders who want to move from theory to production ready implementations.

What you will learn

The curriculum covers core components of modern AI solutions related to RAG:

• What RAG is and why it is needed for working with internal documents, PDFs, notes, and customer data
• Architecture of a retrieval plus generation system and how these parts interact
• Data preparation steps: turning documents into searchable content, handling PDFs and text
• Embeddings, vector databases, and how to structure indexes for efficient search
• How retrieval and generation can be integrated into a single pipeline
• Live demonstrations of building a working RAG system using OpenAI models, with a Telegram chatbot for testing
• Cost considerations and evaluation of answer quality to balance performance and expenses
• How to assemble the pipeline locally, including nuances of local large language models

Format

The course blends short theory sessions with live coding and practical exercises. Attendees will not only understand the theory behind RAG but will also build a basic, working version during the course and learn how to adapt it to their own tasks. By the end, participants will have a ready-to-run example system and a solid understanding of applying these techniques to real-world problems to save time and boost team efficiency.

Agenda and program highlights

• Basics of RAG and why it is necessary for handling real documents
• Architecture of a RAG system and how retrieval and generation work together
• Data preparation for documents, chunking, and embedding strategies
• Infrastructure: vector databases and integration with LLMs
• Practical pipeline construction and hands-on exercises
• Practice 1: Demonstration and testing with OpenAI models and a Telegram bot
• Practice 2: Cost analysis and quality evaluation of answers
• Practice 3: Fully local pipeline construction and working with local LLMs

Speakers and mentors

Natalia Manakova, Senior Data Scientist and AI consultant at SoftServe, PhD, serves as the course mentor. She brings experience in developing AI solutions using LLMs and RAG, with a track record of production ready GenAI/RAG/Agentic systems. Her background includes research in AI and leadership in professional programs, providing guidance on practical approaches to building robust AI systems.

Who should attend

This course targets Data Scientists and ML Engineers (moving from classic ML to GenAI), Data/Business Analysts who work with large text data, AI/GenAI enthusiasts with basic knowledge, and Tech Leads or Architects evaluating AI solutions for production. The content is designed to be applicable across industries where internal documents and knowledge bases are central to decision making.

Format details and logistics

• Date and time: August 25, 26, 27, 28, 29, with sessions starting in the evening Kyiv time on weekdays and morning on Saturday. Each session lasts about 2.5 hours.
• Platform: Zoom, with a link provided the day before the session and accessible on the event page.
• Language: Ukrainian for event delivery; materials and descriptions are in English where indicated for broader accessibility.
• Access: Attendees receive videos, slides, and course materials on the educational platform, along with a certificate of participation upon completion.
• Pricing: Attendee tickets start at 4500 UAH, with various discounts for early registration and group purchases. Payment options include installment plans and discounts for organizations.