RAG Chatbot: Vector Database Approach

In a previous blog, I wrote about a RAG chatbot that I created using a dataset of medical questions and answers. The RAG chatbot embedded the user’s question and retrieved the most relevant answers from an embedding file. Then, using the retrieved answers, GPT-4o-mini generated the final answer. In this blog, we will get rid of the embedding file and use a vector database instead, the code will be repurposed to be smaller in size in order to deploy the chatbot using Docker and AWS.

May 1, 2025 Read
RAG Medical Chatbot

A few years ago, when I started delving deeper into programming, AI was one of the main topics that captivated my interest. It was only natural for me to stumble upon this subject since I was drawn to scientific programming and Python. Back in the day, AI wasn’t the hype as it is today. Most importantly, when it was discussed, it was associated with different domains, not just LLMs as it is today.

April 9, 2025 Read