RAG init Project

A production ready Retrieval Augmented Generation (RAG) application built with Python, LangChain, and ChromaDB. This project demonstrates practical AI engineering skills using completely free LLM APIs.

Overview

This RAG application allows you to:

• Ingest documents into a local vector database
• Retrieve relevant context using semantic search
• Generate answers using free LLM APIs (Groq)

Tech Stack

Component	Technology	Why I Chose It
Framework	LangChain	Standard for RAG, extensive community support
Vector DB	ChromaDB	Free, local first, no external dependencies
LLM	Groq (Llama 3.1)	Free tier, fast inference, open source models
Embeddings	HuggingFace (all-MiniLM-L6-v2)	Free, local, no API calls for embeddings

Architecture

┌─────────────────┐     ┌─────────────────┐     ┌─────────────────┐
│  Source Docs    │────▶│  Text Splitter  │────▶│  HuggingFace    │
│  (data/*.txt)   │     │  (500 char      │     │  Embeddings     │
│                 │     │   chunks)       │     │  (local)        │
└─────────────────┘     └─────────────────┘     └────────┬────────┘
                                                         │
                                                         ▼
┌─────────────────┐     ┌─────────────────┐     ┌─────────────────┐
│  Generated      │◀────│  Groq LLM       │◀────│  ChromaDB       │
│  Answer         │     │  (Llama 3.1)    │     │  (local         │
│                 │     │                 │     │   vector store) │
└─────────────────┘     └─────────────────┘     └─────────────────┘
       ▲                                               ▲
       │                                               │
       └───────────────────────────────────────────────┘
                    User Question

Quick Start

1. Clone and Setup

git clone <repository-url>
cd rag-portfolio
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -r requirements.txt

2. Configure API Keys

cp .env.example .env
# Edit .env and add your free Groq API key
# Get one at: https://console.groq.com (free tier available)

3. Ingest Documents

python ingest.py

4. Run Queries

# Interactive mode
python query.py

# Test mode (for portfolio verification)
python query.py --test

Sample Test Results

Question: What company did the intern work for?
Answer: The intern worked at Wintershall Dea GmbH, one of Europe's 
leading independent natural gas and crude oil producers.

Question: What was the main focus of the internship?
Answer: The internship focused on geological data analysis, reservoir 
characterization, and operational support for North Sea gas production 
facilities over a 12-week period.