🎯 JobHunt: AI-Powered Career Pipeline

JobHunt is a personal automation pipeline designed to eliminate the friction of modern job searching. It programmatically scans job portals, scores your fit with deep AI reasoning, and generates hyper-personalized application packages in seconds.

🚀 Key Features

• Parallel Multi-Key Scoring: High-performance scoring using a pool of multiple API keys (Gemini, Groq, OpenRouter) with automatic rotation and rate-limit handling.
• YC Startup Discovery: Specialized scraper for the Work at a Startup (YCombinator) job board and a curated list of top YC startups.
• Indian Portal Support: Automated scraping for Indian platforms like Instahyre with role and location-based targeting.
• Department-Aware Scanning: Target roles in engineering, data, product, design, sales, or marketing.
• Rich AI Scoring: Every job gets a Letter Grade (A-F), a numerical score, a "Top Project" recommendation, and a "Match Verdict".
• Master Application Workflow: Tailor CV -> Generate PDF -> Find Outreach Targets -> Log to Sheets -> Open Browser.
• Jake's Resume PDF: Generates professional, single-column LaTeX-style resumes (HTML/CSS -> PDF via Playwright).
• Outreach Research: Identifies specific hiring managers and writes personalized LinkedIn/Email messages.
• Visual Application Tracker: Logs everything to a professional Google Sheet with color-coded status badges and grades.

📋 Requirements

• Python 3.10+
• Git
• Playwright (for PDF generation and web scraping)
• API Keys: Gemini (primary), Groq, or OpenRouter (optional for parallel speedup)

🛠️ Quick Start

1. Setup Environment

git clone https://github.com/imHardik1606/JobHunt.git
cd JobHunt
python -m venv venv
.\venv\Scripts\activate
pip install -r requirements.txt
playwright install chromium

2. Configure

1. Create cv.md in the root and paste your full CV in Markdown.
2. Copy .env.example to .env and add your keys.
   • Pro Tip: Add multiple keys like GEMINI_API_KEY_1, GEMINI_API_KEY_2, GROQ_API_KEY_1 to enable high-speed parallel scoring.
3. Update config.py with the companies you want to monitor.
4. Run python main.py sheets_setup for Google Sheets tracking instructions.

3. Run the Pipeline

# Scan for Engineering roles in India, including YC companies
python main.py scan engineering india yc

# View your ranked matches and launch applications
python main.py pipeline

# Or apply directly to a specific job ID
python main.py apply {job_id}

🕹️ Commands

🔍 Find Jobs

Command	What it does	Example
scan	Scan portals by dept + level + location	python main.py scan engineering fresher yc
internships	YC + WaaS internships only	python main.py internships india
pipeline	View ranked scored matches	python main.py pipeline 7

📨 Apply

Command	What it does	Example
apply	Full apply workflow	python main.py apply 3
outreach	Standalone outreach research	python main.py outreach "Razorpay Backend"

📊 Track

Command	What it does	Example
status	Pipeline stats	python main.py status
sheets_setup	Google Sheets setup guide	python main.py sheets_setup

Scan Flags

A table showing all available flags that can be combined:

Flag	Options	Default	Example
Department	engineering, data, product, design, sales, marketing	engineering	scan data
Experience	fresher, intern, junior, mid, senior, any	fresher	scan engineering intern
Location	india, remote	none (all)	scan engineering india
Portal	yc	off	scan engineering yc

💡 Example Combinations

• python main.py scan engineering fresher yc
• python main.py scan engineering intern india yc
• python main.py scan data fresher remote
• python main.py internships india
• python main.py scan engineering any

🔄 Workflow

graph TD
    subgraph Discovery ["1. Job Discovery"]
        A1[Greenhouse/Lever/Ashby APIs] --> B
        A2[YC Work at a Startup] --> B
        A3[Instahyre Scraper] --> B
    end

    B{Deduplication & Filter} -- New Relevant Jobs --> C

    subgraph Scoring ["2. Parallel AI Scoring"]
        C[KeyPool Manager] --> D1[Gemini 1.5 Flash]
        C[KeyPool Manager] --> D2[Groq Llama 3]
        C[KeyPool Manager] --> D3[OpenRouter Free]
        D1 & D2 & D3 --> E[Match Result: Grade, Score, Verdict]
    end

    E --> F[(Local Database)]

    subgraph Workflow ["3. Application Workflow"]
        F --> G[python main.py pipeline]
        G -- Select Job --> H[Resume Tailoring]
        H --> I[PDF Generation]
        I --> J[Outreach Research]
        J --> K[Log to Google Sheets]
        K --> L[Launch Browser]
    end

Loading

🏗️ Architecture

High-Speed Scoring Engine

JobHunt uses a KeyPool architecture to distribute scoring tasks across multiple AI models:

• Parallel Workers: Automatically calculates the optimal number of workers based on your available API keys.
• Failover Logic: If Gemini hits a rate limit, the system automatically rotates to Groq or another Gemini key.
• Model Support: gemini-1.5-flash, llama3-8b (via Groq), and openrouter/free meta-models.

Discovery Pipeline

The discovery engine combines direct portal API calls with browser-based scraping:

• Direct: Greenhouse, Lever, and Ashby APIs for 100+ top companies.
• Scraped: YCombinator (Work at a Startup) and Instahyre via Playwright.
• Deduplication: Automatic global deduplication by URL across all sources.

🤝 Customization

To monitor a new company, add it to COMPANIES in config.py:

{"name": "Anthropic", "portal": "greenhouse", "id": "anthropic"}

Support portals: greenhouse, lever, ashby.

💰 Cost

• Job Discovery: ₹0 (Public APIs & Playwright)
• AI Scoring: ₹0 (Gemini/Groq/OpenRouter Free Tiers)
• Resume PDF: ₹0 (Playwright)
• Total: ₹0

Important

The parallel scoring engine is extremely fast. Ensure you have enough API keys in .env to sustain high worker counts.

Project Summary

JobHunt is a high-performance automation engine designed to eliminate the manual grind of job seeking. It doesn't just find jobs; it acts as a full-stack career assistant by:

• Automated Discovery: Real-time scanning of YC (Work at a Startup), Instahyre, Greenhouse, Lever, and Ashby.
• Deep AI Reasoning: Using a parallelized "KeyPool" (Gemini, Groq, OpenRouter) to grade your fit (A-F) against your Markdown CV.
• Surgical Tailoring: Generating hyper-personalized, single-column LaTeX-style resumes (Jake's Template) via HTML/CSS & Playwright.
• Outreach Intelligence: Identifying hiring managers and generating personalized LinkedIn/Email templates.
• Centralized Tracking: Automatically logging every application to a color-coded Google Sheets dashboard.