Nightly LinkedIn scrape that summarises listings through the Hugging Face API and pushes them to a GitHub repo. First real project I built, back in January 2024.
Wanted to see if I could search LinkedIn, get AI summaries of postings, and dump the results into a GitHub repo without touching anything manually.
Solo. Built in January 2024, before I'd properly learned to code, mostly by chaining tutorials together.
You run main.py, type in a job title like "Software Engineer Intern", and the rest happens on its own. scraper.py drives LinkedIn through Selenium and pulls back recent listings. Each one gets fed to the Hugging Face API through api_handler.py, which runs facebook/bart-large-cnn for a short summary.
The output writes locally to jobs.json and job_log.md, then github_uploader.py pushes it to a GitHub repo through PyGithub. Ends up looking like a self-updating job tracker.
MIT licensed and public. I'd describe it as super basic and very messy, and I gave up on Python not long after finishing it. Rediscovered the repo on an old laptop and archived it as it was rather than cleaning it up. First time I'd chained multiple APIs together into something that actually did what I wanted.