Hello! I'm Anmol Ratan Srivastava, currently pursuing a Bachelor's in Computer Science Engineering with a specialization in Artificial Intelligence and Machine Learning. This repository contains my submission for the Gen AI Engineer / Machine Learning Engineer internship assessment, where I developed a Retrieval-Augmented Generation (RAG) Model and an Interactive QA Bot Interface.
This project is divided into two parts:
The RAG model is designed to answer questions based on a provided dataset or document. It uses a vector database like Pinecone to store and retrieve document embeddings efficiently and a generative model (e.g., Cohere API) to generate accurate, context-aware answers.
I built an interactive frontend using Gradio, allowing users to upload documents (e.g., PDFs) and ask questions in real-time. The interface processes the uploaded content using the RAG model, retrieving relevant document sections and generating real-time responses.
- Document Upload & Real-Time QA: Users can upload documents and ask questions based on the content.
- Efficient Retrieval: Uses a vector database for fast and accurate document embedding retrieval.
- Generative Model: Answers are generated contextually using a state-of-the-art generative model.
- Scalable and Modular Code: Designed for easy scalability and modularity.
- Colab: Use the Colab notebook to explore the end-to-end RAG pipeline.
- Gradio: Try the interactive QA bot by uploading documents and asking questions in real-time via the Gradio interface.
The project is containerized using Docker, making it easy to deploy locally or in the cloud. Detailed deployment instructions can be found in the repository.