Cold Email Generator

🎯 Project Overview

The Cold Email Generator is an intelligent tool designed specifically for service companies to automate and personalize their business outreach efforts. By leveraging advanced AI technologies, it extracts job listings from company career pages and generates highly targeted cold emails that match specific job requirements with relevant service offerings.

Llama 3.1 LLM Groq LangChain Streamlit ChromaDB Python

⚡ How It Works

Live Demo Interface

URL Input

→

Web Scraping

→

Job Extraction

→

Resume Matching

→

Email Generation

🔧 Technical Workflow Deep Dive

Data Extraction Layer

Web Scraping Engine: Uses BeautifulSoup4 and Selenium to extract job postings from career pages. Handles dynamic content loading and JavaScript-rendered pages.

# Web scraping implementation
from selenium import webdriver
from bs4 import BeautifulSoup
import requests

def extract_jobs(url):
    driver = webdriver.Chrome()
    driver.get(url)
    html = driver.page_source
    soup = BeautifulSoup(html, 'html.parser')
    return parse_job_listings(soup)

Job Parsing & Structuring

LangChain Integration: Processes raw HTML content to extract structured job data including titles, requirements, descriptions, and company information.

# Job parsing with LangChain
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.schema import Document

def structure_job_data(raw_jobs):
    structured_jobs = []
    for job in raw_jobs:
        job_doc = Document(
            page_content=job['description'],
            metadata={
                'title': job['title'],
                'requirements': job['requirements'],
                'company': job['company']
            }
        )
        structured_jobs.append(job_doc)
    return structured_jobs

Vector Database Storage

ChromaDB Integration: Converts job descriptions into embeddings and stores them in a vector database for semantic similarity matching.

# ChromaDB vector storage
import chromadb
from langchain.vectorstores import Chroma
from langchain.embeddings import HuggingFaceEmbeddings

def store_in_vectordb(job_documents):
    embeddings = HuggingFaceEmbeddings()
    vectorstore = Chroma.from_documents(
        documents=job_documents,
        embedding=embeddings,
        persist_directory="./job_vectors"
    )
    return vectorstore

Resume Portfolio Matching

Semantic Search: Matches job requirements with relevant portfolios/resumes using cosine similarity. Identifies best-fit candidates for each position.

# Resume matching algorithm
def match_resumes_to_jobs(job_embeddings, resume_portfolio):
    matches = []
    for job in job_embeddings:
        similar_resumes = vectorstore.similarity_search(
            job.page_content,
            k=3,
            filter={'skills': job.metadata['requirements']}
        )
        matches.append({
            'job': job,
            'matched_resumes': similar_resumes
        })
    return matches

AI Email Generation

Llama 3.1 via Groq: Generates personalized cold emails using the matched job-resume pairs. Creates contextually relevant content with specific value propositions.

# Email generation with Llama 3.1
from groq import Groq

def generate_cold_email(job_match, company_info):
    client = Groq(api_key="your-groq-api-key")
    prompt = f"""
    Generate a professional cold email for:
    Job: {job_match['job'].metadata['title']}
    Company: {company_info['name']}
    Matched Resume: {job_match['matched_resumes'][0]}
    
    Focus on specific value propositions and relevance.
    """
    
    response = client.chat.completions.create(
        messages=[{"role": "user", "content": prompt}],
        model="llama3-70b-8192"
    )
    return response.choices[0].message.content

Output Delivery

Streamlit Interface: Presents generated emails in an intuitive web interface with editing capabilities, export options, and campaign management features.

# Streamlit interface
import streamlit as st

def display_generated_emails(emails):
    st.title("Generated Cold Emails")
    for i, email in enumerate(emails):
        with st.expander(f"Email {i+1}: {email['subject']}"):
            st.text_area("Email Content", email['content'])
            if st.button(f"Export Email {i+1}"):
                export_email(email)

📧 Complete Example

                Scenario: A web development agency wants to reach out to Nike for their Senior Frontend Developer position
            

Input Data:

Job Title: Senior Frontend Developer
Company: Nike
Requirements: React, TypeScript, 5+ years experience, E-commerce
Location: Portland, OR
Posted: 2 days ago

Matched Resume: John Smith - Frontend Specialist
- 6 years React development
- E-commerce platform expertise
- TypeScript proficiency
- Portfolio: 15+ successful projects

Generated Cold Email:

Subject: Enhancing Nike's Frontend Development with Proven E-commerce Expertise Dear Nike Hiring Team, I noticed your recent posting for a Senior Frontend Developer position and wanted to reach out regarding our specialized frontend development services. Our team has extensive experience in the exact technologies you're seeking: • **React & TypeScript Mastery**: Our lead developer, John Smith, brings 6+ years of React development experience with deep TypeScript expertise • **E-commerce Specialization**: We've successfully delivered 15+ e-commerce platforms, understanding the unique challenges of high-traffic retail environments • **Performance Focus**: Our solutions consistently achieve 95+ PageSpeed scores, crucial for conversion optimization Given Nike's position as a global e-commerce leader, we understand the critical importance of: - Seamless user experiences across devices - Fast loading times for product catalogs - Scalable architecture for traffic spikes during launches **Value Proposition**: Instead of the typical 3-6 month hiring and onboarding process, we can provide immediate frontend development support while you find the perfect full-time candidate. **Recent Success**: We recently helped [Similar Company] reduce their product page load time by 40%, resulting in a 15% increase in conversions. Would you be open to a brief 15-minute call to discuss how our frontend expertise could support Nike's immediate development needs? Best regards, [Your Name] [Company Name] [Contact Information] P.S. I've attached our portfolio showcasing similar e-commerce projects that might interest your team.

Email Analytics:

Personalization Score: 92/100
Relevance Match: 89/100
Professional Tone: 95/100
Call-to-Action Strength: 88/100

Key Matching Elements:
✓ Technology stack alignment (React, TypeScript)
✓ Industry experience (E-commerce)
✓ Seniority level match (6+ years vs 5+ required)
✓ Location consideration (Remote-friendly)
✓ Portfolio relevance (15+ projects)

🏗️ System Architecture

Streamlit Frontend

↕

LangChain Processing

↕

ChromaDB Vector Store

↕

Groq + Llama 3.1

Data Flow:

Input Processing

Career page URLs → Web scraper → Raw HTML → Job parser → Structured data

Vector Processing

Job descriptions → Embeddings → ChromaDB storage → Similarity search → Resume matching

Email Generation

Matched pairs → Prompt engineering → Llama 3.1 → Generated emails → Quality scoring

Output Delivery

Email content → Streamlit interface → User review → Export/Send → Campaign tracking

🚀 Key Features

Intelligent Matching

Uses vector similarity to match job requirements with portfolio skills, ensuring highly relevant email content and improved response rates.

Dynamic Personalization

Each email is uniquely crafted based on specific job requirements, company culture, and matched candidate profiles.

Batch Processing

Process multiple job listings simultaneously, generating dozens of targeted emails in minutes rather than hours.

Quality Scoring

Built-in analytics rate email quality, relevance, and personalization level to optimize outreach effectiveness.