Job Hunt Assistant

A multi-agent AI system that finds jobs, analyzes requirements, and generates tailored application materials

Overview

Job Hunt Assistant is a multi-agent system that automates the entire job search process using CrewAI, LangChain, and Python. This portfolio-worthy project demonstrates practical AI engineering skills while helping you save time on repetitive job hunting tasks.

The job market is competitive, and nearly every employer values AI skills. Why not let AI help with your job hunt? This system handles everything from analyzing job descriptions to drafting personalized LinkedIn outreach messages, freeing you up for what matters most: interview prep, upskilling, and getting hired.

---

Technologies

• Python
• Google Gemini – LLM for content generation
• Streamlit – Web interface
• LangChain – Agent framework and tools
• CrewAI – Multi-agent orchestration

---

Features

Core Functionality

• Job Discovery - Browse relevant positions from USAJobs API
• Smart Analysis - AI agents analyze job descriptions and requirements
• Content Generation - Create customized cover letters and resumes
• LinkedIn Outreach - Draft personalized connection messages
• Application Tracking - Persistent logging of all applications

Architecture

flowchart TD
    USER[👤 User] --> UI[Streamlit UI]
    UI --> API[🌐 USAJobs API]
    UI --> CREW[CrewAI Orchestrator]

    CREW --> A1[📋 JD Analyst Agent]
    CREW --> A2[✍️ Resume & CL Agent]
    CREW --> A3[💼 Messaging Agent]

    A1 & A2 & A3 --> LLM[🧠 Gemini LLM]
    LLM --> TRACKER[💰 Token Tracker]

    A2 --> OUT1[📄 Resume & Cover Letter]
    A3 --> OUT2[💬 LinkedIn Messages]

Agent	Role
JD Analyst Agent	Analyzes job descriptions and extracts key requirements & skills
Resume & CL Agent	Generates tailored resume summaries and cover letters
Messaging Agent	Drafts personalized LinkedIn outreach messages

---

Cost Tracking & Budget Alerts

The application includes comprehensive cost monitoring to help manage API expenses.

• Per-job cost tracking - See exactly what each application costs
• Session totals - Track cumulative costs in the sidebar
• Lifetime statistics - Persistent tracking across sessions
• Budget alerts - Warnings when approaching/exceeding limits
• Model breakdown - Costs split by Gemini model used

How It Works

1. Token Extraction: When CrewAI runs the pipeline, token usage is extracted from the result:

   • prompt_tokens - Input tokens sent to the model
   • completion_tokens - Output tokens generated
   • total_tokens - Combined count

2. Cost Calculation: Costs are calculated using Gemini API pricing:

   Model	Input (per 1M)	Output (per 1M)
   gemini-2.0-flash	$0.10	$0.40
   gemini-2.5-flash	$0.30	$2.50
   gemini-1.5-pro	$1.25	$5.00

3. Budget Monitoring: Set a budget limit in .env:

   BUDGET_LIMIT=1.0  # USD per session
   

   The sidebar displays usage percentage and warns when exceeded.

4. Data Persistence: Usage data is saved to data/usage_log.json with:

   • Per-run breakdowns (tokens, cost, timestamp)
   • Lifetime totals across all sessions

---

Installation

Prerequisites

Install uv:

curl -LsSf https://astral.sh/uv/install.sh | sh

Setup

# Clone the repository
git clone https://github.com/sheygs/job-hunt-assistant.git
cd job-hunt-assistant

# Create a virtual environment
uv venv
source .venv/bin/activate

# Install dependencies
uv pip install -r requirements.txt

# Set up environment variables
cp .env.example .env

Configure API Keys

The application requires two API keys:

1. USAJobs API Key

1. Visit the USAJobs Developer Portal
2. Navigate to the API Request Page
3. Fill out the request form with your details
4. You'll receive your API key via email

2. Gemini API Key

1. Visit Google AI Studio
2. Sign in with your Google account
3. Navigate to "Get API Key" and create a new key

3. Update Your .env File

# Open .env and add your keys
USAJOBS_API_KEY = "your_usajobs_api_key"
GEMINI_API_KEY = "your_gemini_api_key"
EMAIL_ADDRESS = "your_email_address"
BASE_URL = "your_search_base_url"
...

---

Usage

Start the Streamlit Interface

streamlit run app.py

Configure Your Preferences

1. Enter your job search criteria (role, location, keywords)
2. Upload your base resume and profile information
3. Select job listings to target
4. Let the agents generate customized materials

Review and Apply

1. Review generated cover letters and resumes
2. Check drafted LinkedIn outreach messages
3. Edit as needed before sending
4. Track applications in the logging dashboard

---

Project Module

job-hunt-assistant/
│
├── app.py                    # Main Streamlit entry point
│
├── core/                     # Pipeline orchestration
│   └── orchestrator.py       # CrewAI sequential pipeline
│
├── agents/                   # AI agent definitions
│   ├── jd_analyst.py         # Job description analyzer
│   ├── resume_cl_agent.py    # Resume & cover letter writer
│   └── messaging_agent.py    # LinkedIn outreach writer
│
├── services/                 # Business logic layer
│   ├── job_service.py        # Job processing & tracking
│   ├── jobs_api.py           # USAJobs API client
│   └── storage_service.py    # Session state management
│
├── ui/                       # Streamlit UI components
│   ├── search_tab.py         # Job search interface
│   ├── history_tab.py        # Application history
│   ├── settings_tab.py       # Settings configuration
│   └── components.py         # Reusable UI components
│
├── utils/                    # Utilities
│   ├── token_tracker.py      # Cost tracking & budget alerts
│   ├── llm.py                # Gemini LLM factory
│   ├── config.py             # Configuration management
│   ├── constants.py          # Pricing & constants
│   ├── retry.py              # Retry with backoff
│   └── logger.py             # Logging setup
│
├── data/                     # Data storage
│   ├── applications_log.csv  # Application history
│   ├── usage_log.json        # Token usage tracking
│   └── cover_letters/        # Generated cover letters
│
├── logs/                     # Application logs
├── requirements.txt
├── .env.example
└── README.md

---

Future Enhancements

• Add support for LinkedIn, Indeed, and other job boards
• Implement automated application submission
• Build email monitoring for interview callbacks
• Add resume parsing from existing PDFs
• Create analytics dashboard for application tracking

---

Notes

Important: This tool is designed to assist and augment your job search, not replace human judgment. Always review and personalize generated content before sending applications. Use this as a learning project to develop your AI engineering skills while making your job hunt more efficient.

---

License

This project is licensed under the MIT License - see the LICENSE file for details.

---

Acknowledgments

• CrewAI - Multi-agent framework
• LangChain - LLM application framework
• Streamlit - Web app framework
• Google Gemini - LLM provider
• USAJobs API - Job listings API