RAG System

I built a fully functional Retrieval-Augmented Generation (RAG) system to deeply understand how modern AI systems work in real-world production environments.

The system allows users to upload documents, automatically processes them (chunking + embeddings), stores them in a vector database, and retrieves the most relevant context when answering user queries using a language model.

Core Features

Document upload and processing pipeline

Text chunking and embedding generation

Vector similarity search

Context-aware LLM responses

Background task handling

Local model support (Ollama)

Production-ready deployment setup

Technical Stack & Architecture

Backend:

FastAPI for high-performance async APIs

MVC architecture for clean code organization

Background processing using Celery

Local LLM integration via Ollama

Data & Retrieval:

Embeddings generation

Vector database for similarity search

RAG pipeline orchestration

DevOps & Deployment:

*Docker containerization

CI/CD with GitHub Actions

Production-ready deployment workflow

Environment-based configuration