RAG-Based Cold Email Generation System using LLMs

🚀 Overview

This project is an AI-powered Cold Email Generator that automates the creation of personalized outreach emails based on job postings.

This project implements a retrieval-augmented cold email generation system that extracts job requirements from career pages and generates context-aware outreach emails. It leverages a vector database (ChromaDB) to retrieve relevant portfolio links, enabling personalized and targeted email generation using LLMs

💡 Key Features

• Retrieval-Augmented Personalization (RAG)
• Semantic Portfolio Matching via Vector Search
• Dynamic Prompt Construction using LangChain
• Automated Job Parsing from Web Pages
• Low-Latency Inference using Groq API

🧠 System Architecture

User Input (Careers Page URL)
↓
Web Scraping (Job Description Extraction)
↓
LLM Parsing (Role, Skills, Requirements)
↓
Vector DB Retrieval (ChromaDB - Portfolio Matching)
↓
Prompt Construction (LangChain)
↓
LLM (LLaMA3 via Groq)
↓
Generated Cold Email
↓
Streamlit UI

🔗 How It Works

1. Extract job data from careers page using web scraping
2. Parse and structure job requirements using LLM
3. Retrieve relevant portfolio links via semantic similarity search (ChromaDB)
4. Construct context-aware prompts using LangChain
5. Generate personalized cold emails using LLaMA3 (Groq API)

🏗️ Tech Stack

Languages: Python
LLM & Orchestration: LangChain, LLaMA3 (Groq API)
Vector Database: ChromaDB
Data Processing: WebBaseLoader (Web Scraping)
Frontend: Streamlit

🚀 Future Improvements

• Multi-page job extraction and parsing
• Advanced RAG (re-ranking, hybrid search)
• Email tone customization (formal, casual, recruiter-specific)
• Backend API deployment using FastAPI