Job Scraper

A Python and Streamlit app that scrapes LinkedIn job postings and uses an LLM to rank them against your resume, producing a shortlist sorted by how well each job fits your background.

Status: Early personal project. Built when GPT-3.5 first came out, as my first time working with an LLM in my own project. It works end to end but is rough around the edges and not production-ready.

What it does

• Scrapes LinkedIn job postings by position, location, and experience level using Selenium and the linkedin_jobs_scraper library
• Accepts a resume upload (PDF), with text extracted via pdfplumber
• Sends the resume and scraped jobs to an LLM, which ranks the postings from most to least relevant and removes duplicates
• Runs through a simple Streamlit interface

Built with

• Python
• Streamlit (UI and resume upload)
• Selenium and linkedin_jobs_scraper (job scraping)
• pdfplumber (resume text extraction)
• LLM API via the OpenAI client, pointed at the Perplexity endpoint (relevance ranking)

Getting started

1. Install dependencies:

   pip install -r requirements.txt

2. Create a .env file in the project root:

   GENAI_API_KEY=your_api_key_here
   MODEL=your_model_name_here
   

3. Make sure you have a matching ChromeDriver (chromedriver.exe) for your installed version of Chrome.

4. Run the app:

   streamlit run main.py

5. In the browser, upload your resume, enter a position and location, choose an experience level, and click Search. Log in to LinkedIn in the window that opens; scraping begins once you reach your feed.

Notes and limitations

• This is an early experiment, not a polished or production-ready tool.
• LinkedIn scraping can break when the site changes and may be subject to LinkedIn's terms of service. Use responsibly and at your own risk.
• Requires a manual LinkedIn login during the run.
• Scraped results are also written to a local jobs.txt file.

License

© 2024 Vincent Wirawan. All rights reserved.