RAG Demystified (For Student and beginners)

​About event

​AI systems can answer instantly and confidently while still getting things completely wrong. This workshop explores why that happens and how Retrieval-Augmented Generation (RAG) helps AI look things up before answering. Through real-world stories, interactive demos, and a guided hands-on build, you'll see how reliable AI systems are designed around information they can actually retrieve and verify.

​What we'll get into

​The session will be practical and implementation-focused. Expect to explore:

• ​Why AI can produce confident but incorrect answers

• ​Retrieval-Augmented Generation (RAG) and how retrieval grounds AI responses

• ​Embeddings, chunking, and vector search

• ​Hybrid search using BM25 and dense retrieval

• ​Reranking, Contextual Retrieval, and RAG evaluation

• ​When to use RAG, fine-tuning, or long-context approaches

​What we'll build

• ​A RAG system over documents brought by participants

• ​A workflow that retrieves relevant information before generating an answer

• ​An experience of testing the system with questions and seeing how retrieval affects its answers

​Who should come?

​This session is for students and beginner-level folks who want to understand how reliable AI systems are built, whether you're new to AI, experimenting with chatbots, or working on an AI project.

​What you'll need

​Please bring your own laptop.

​No prior software installation is required.

​Facilitated by

​K. Akshay Kumar, Kovan Labs.