KalyanM45/CareerCrawl ? reverse-engineered prompt

Build me a Python tool that crawls company career pages like Workday and Eightfold, pulls the latest job listings, and keeps only real tech roles like software, data, and engineering.

I want it to use an LLM to read each posting, figure out whether it is actually a tech job, and extract useful details like experience needed, work mode, role type, location, and the original link. It should skip obvious non tech roles, remember jobs it has already seen, and black list anything that gets classified as irrelevant so it never shows up again.

The output should be a clean feed that updates as it runs, saved to both JSON and CSV, with jobs sorted newest first. If you can, also make a simple way to view the results in the browser so I can quickly scan the curated list. Use the Groq API key from the environment, and look up current docs online if you need to.