Skip to main content

๐Ÿš€ Run Job Explorer via Streamlit

Want to try the app right now without setting up anything locally? This page shows you how.


๐Ÿ”— Open the Appโ€‹

Visit the hosted version of Job Explorer:

๐Ÿ‘‰ streamlit.app/job-explorer (Replace with actual URL)

No login or installation needed โ€” just open the link and start using it.


๐Ÿงญ What Youโ€™ll Seeโ€‹

Once loaded, the app walks you through a guided job search pipeline with multiple tabs:

1. Query Tabโ€‹

  • Enter a job title or keywords (e.g., "data scientist")
  • Set the number of results and region if supported
  • Click "Fetch Jobs" to start the process

2. Config Tabโ€‹

  • Tweak advanced settings (optional)
    • Scoring weights
    • Prompt logic
    • Output structure
  • Defaults work well, but power users can tune it here

3. Results Tabโ€‹

  • View scored job listings
  • Download final results as .csv
  • Explore by category or confidence

4. Prompt Tabsโ€‹

  • Jinja Editor: Customize prompt templates used for LLM ranking
  • JSON Editor: Manually adjust job inputs or configuration schemas

๐Ÿงช What Happens Behind the Scenesโ€‹

When you click Fetch Jobs, the following pipeline is triggered:

  1. Job query hits online job boards (via scraper)
  2. Results are parsed and filtered
  3. Each job is scored using an LLM prompt (based on your resume)
  4. Results are saved and displayed

Youโ€™ll see live progress in the sidebar as each step completes.


๐Ÿ“ฆ Outputs You Getโ€‹

  • results.csv: All jobs with ranking and metadata
  • Optional: Intermediate .jsonl files (visible if enabled)
  • Prompt previews: Let you see the raw model input/output

๐Ÿ›Ÿ Having Issues?โ€‹

If the hosted app stalls or errors:

  • Refresh the page
  • Reduce number of results in the query tab
  • Use the Run Locally) version for full control

Ready to try advanced features or offline usage?
Continue to Run Locally โ†’