Mahashwetha Resume Tailor
Generate a per-company tailored resume using Gemini 2.5 Flash via resume_tailor.py.
How to run
For a single job, run resume_tailor.py with the job posting URL and company name as arguments.
For batch mode (all companies with status done and non-LinkedIn links in the tracker), run it with no arguments.
Output
Saved to the resume_adjusted/ folder configured in config.py, named resume_{CompanyName}.docx. The script skips a company if the file already exists — delete the file to regenerate it.
What it does
Fetches the job description from the URL, reads the base resume, and asks Gemini to suggest minimal tweaks — reordering skills, surfacing existing keywords. It never fabricates experience. Gemini's markdown formatting (**bold**) is stripped automatically before writing to the DOCX.
Site compatibility
Most standard job pages work. Workday works via JSON-LD extraction. Ashby and internal portals (like BPCE) may fail — in that case, skip the URL and tailor manually, or paste the job description text into a temp file and adjust the call.
LinkedIn URLs are skipped automatically in batch mode.
If you hit rate limits
The free Gemini tier occasionally returns 429 errors — the script retries automatically. If it keeps failing, wait a minute and re-run. Use gemini-2.5-flash, not gemini-2.0-flash (that quota runs out faster).
Note: on-demand only
Resume tailor is NOT part of the daily job search run — controlled by RESUME_TAILOR_ENABLED in config.py (default: False).
When the user asks to tailor a resume:
- Set
RESUME_TAILOR_ENABLED = Trueinconfig.py - Run the tailor
- Set
RESUME_TAILOR_ENABLED = Falseinconfig.pyagain