Elena Kor Blog in Medium
As a developer I’ve worked with generative AI on a few projects before. And this time around, I was tasked with setting up my very first RAG project — for Intellyx — industry analysis, advisory, and training firm focused on digital transformation for business executives and IT professionals. The Intellyx team has created and published a wealth of valuable content over the years. It’s an exciting opportunity to build an app that integrates all that knowledge into a RAG store, allowing people to interact with it directly through chat. Here’s a look at my experience, the tech I used, and a few key takeaways from the project.
The idea was to create a custom AI chat system that allows users to interact with Intellyx analysts, either individually or as a company, and retrieve relevant information from the RAG store. The system needed to be able to filter the information based on the analysts’ contributions, ensuring that the responses were tailored to what each analyst had written or published. When interacting with “All of Intellyx,” the chat should provide the most relevant analyst response first, as well as respond to questions regarding the general company information, such as subscription options and pricing. Additionally, chat responses needed to include sources, complete with the article links and age (how long ago the article was published). Given the fast pace of the tech world, users need to know how current the content is to gauge its relevance…
Read the article on Medium here: https://medium.com/@elena_kor/building-my-first-rag-system-a-full-stack-developers-journey-with-ai-node-js-and-supabase-d11e6581d36a
