Insights from One Year of Customer Implementations of RAG-based Chatbot Systems Using MongoDB Vector Search

Colloquium of the Responsible Digital Society Research Network

Speaker:

Philip Eschenbacher, MongoDB

Abstract:

In this talk we will explore the journey of building chatbots using Retrieval-Augmented Generation (RAG) systems with MongoDB Atlas, focusing on valuable insights and lessons learned over the past year. As RAG systems become increasingly vital for enhancing information retrieval and generation and to tailor chatbots to corporate documents, our experiences and customer implementations provide a rich source of knowledge.
We will delve into the technical aspects and best practices on MongoDB Vector Search and how to make customer data accessible for language models. Through real-world examples, we will highlight the challenges faced and solutions devised by our customers in building such chatbots, showcasing the practical applications and benefits.

Curriculum Vitae:

Philip's tech journey began as a developer, with a keen interest sparked by programming on a Commodore 64 at age 10. By 1998, he emerged as a product manager for a pioneering Swiss Internet banking solution. His career path included significant detours into operations, notably in the large-scale migration from Tru64 to Linux, and a thrilling stint at a startup refining his web development skills. Most recently, before joining MongoDB, Philip spearheaded the development of an innovative web-based workplace solution for a large Swiss bank, transitioning it from MySQL to MongoDB. In this role, he demonstrated his versatility, wearing multiple hats from developer to solutions architect and team leader, before taking on his current position as an Advisory Solutions Architect at 🍃MongoDB.

Veranstaltung übernehmen als iCal