Mór Kapronczay, lead ML engineer at Superlinked, gave an eye-opening talk at the 2024 Data Science Festival, about using Retrieval-Augmented Generation (RAG) to create better chatbots for HR applications.

Mór explained how RAG bridges the gap between retrieval and generation, making chatbots smarter and more accurate—key for handling HR questions where precision matters. While most people focus on improving the text that AI generates, Mór highlighted that real performance gains often come from optimizing how data is retrieved, especially for complex datasets like HR policies and employee info.

He also shared cool techniques for combining text and numeric data into a single vector, allowing chatbots to handle mixed data seamlessly. To top it off, he demonstrated how to build a high-performing RAG chatbot specifically for HR inquiries, showing how it can save teams time and improve employee access to information.

Want to give it a try? Check out the VectorHub article for tips and grab the tools from the GitHub repo. And if you missed it, you can catch the full talk below!

‍