Opinion on Gen AI?

It depends on what kind of use cases you are targetting

One of the features I built in a recent project was building a RAG bot so that a user could ask about stuff from the core applications under specific groups. Under the hood was a RAG which was utilizing SQL query agent. Additionally for each group, you could so upload a PDF or text file (like internal documents) in order to give more context on that topic.

Simply being an AI engineer wouldn't cut it. We had to involve pgvector extension for Postgresql to involve similarity searching, Redis semantic cache to save API calls for similar API calls, event driven architecture using Kafka to ensure a scalable load. Furthermore, we had to assign rag_user user to postgres to ensure it only has read only rights.

By passing the user details as part of the query along with conversation history, we also ensured a personalised and Conversational RAG bot

Don’t understand any of these comments 🫣