Intelligent job search automation with company ratings integration
A personal tool that automatically scrapes job listings from LinkedIn and Indeed, enriches them with Kununu company ratings, and provides an interactive dashboard for efficient job hunting.
BEWARE: This is a work in progress to test a number of assumptions. Everything can fail all the time.
Company Rating Scraper automates the tedious process of job hunting by:
- Collecting job postings from multiple sources (LinkedIn, Indeed)
- Fetching company reputation data from Kununu
- Providing intelligent filtering and blacklisting
- Offering a clean, interactive dashboard for review
Built with a clean hexagonal architecture for maintainability and extensibility.
- Multi-Source Job Scraping - Automated collection from LinkedIn and Indeed via JobSpy
- Company Ratings - Enriches jobs with Kununu ratings and review counts
- Interactive Dashboard - Vue 3 + TypeScript frontend for the Company Rating system.
- Smart Blacklisting - Hide unwanted companies or individual jobs
- Alternative Names - Handle companies with different names across platforms
- File-Based Storage - Simple JSONL format, no database required
- Clean Architecture - Ports & adapters pattern for easy testing and extension
- Python 3.12 or higher
- uv - Fast Python package manager
git clone <repository-url>
cd CompanyRating
# Install dependencies
uv sync
# Install Playwright browsers (required for Kununu scraping)
uv run playwright install chromiummake docker-upThis will:
- Build both frontend and backend images
- Start containers with proper networking
- Mount data directory for persistence
- Expose services on localhost
- Frontend: http://localhost:8080
- Backend API: http://localhost:8000
- API Docs: http://localhost:8000/docs
1. Scrape Jobs
Run the daily scraping routine to fetch new jobs:
uv run src/main.pyThis performs:
- Job scraping from LinkedIn and Indeed
- Blacklist filtering
- Deduplication
- Data persistence to
data/jobs.jsonl
Some companies appear with different names on job boards vs. rating sites:
# 1. Export companies without ratings
# Edit src/main.py: task = Tasks.EXPORT
uv run src/main.py
# 2. Edit data/companies_missing_ratings.csv
# Fill alternative_name column with correct Kununu names
# Use "void" for companies not on Kununu
# 3. Import and re-scrape
# Edit src/main.py: task = Tasks.IMPORT
uv run src/main.pyEdit src/infrastructure/scrapers/jobspy_scraper.py:
search_terms = [
"JOB TITLE 1",
"Fullstack JOB TITLE 2",
"AI JOB TITLE 3",
"...",
]
location = "YOURPLACEOFWORK"
hours_old = 24 * 7 # Last 7 days
results_wanted = 100Edit src/main.py to change the execution task:
task = Tasks.DAILY # Daily job scraping
task = Tasks.EXPORT # Export companies missing ratings
task = Tasks.IMPORT # Import alternative names and re-scrapesrc/
├── domain/ # Core business logic
│ ├── entities/ # Job, Company entities
│ ├── ports/ # Interface definitions
│ └── value_objects/ # Immutable value objects
├── infrastructure/ # External integrations
│ ├── adapters/ # File storage, blacklist management
│ └── scrapers/ # Kununu (Playwright), JobSpy
├── application/ # Use cases
│ └── services/ # JobOrchestrator
├── presentation/ # User interfaces
│ └── streamlit_app.py # Interactive dashboard
└── main.py # CLI entry point
All data stored in the data/ directory:
| File | Purpose |
|---|---|
companies.jsonl |
Company data with ratings |
jobs.jsonl |
Scraped job listings |
company_blacklist.txt |
Hidden companies |
job_blacklist.txt |
Hidden job IDs |
companies_missing_ratings.csv |
Manual name correction workflow |
filtered_*.csv |
Exported results |
Built with Hexagonal Architecture (Ports & Adapters):
- Domain Layer: Pure business logic, framework-agnostic
- Application Layer: Use cases and orchestration
- Infrastructure Layer: External concerns (file I/O, web scraping)
- Presentation Layer: User interfaces (Streamlit, CLI)
Benefits:
- Easy to test
- Swappable implementations (file storage → database)
- Clear separation of concerns
- Rate Limiting: Kununu scraper includes 1-5s delays for respectful scraping
- Stealth Mode: Playwright uses anti-detection techniques
- German Market: Optimized for German job market (Indeed DE, Kununu)
- Privacy: All data stored locally, no external services required