Run a morning job scan that scores roles and tailors resumes on request

Kalita, a data engineer, built a job search pipeline for a senior-level friend inside the Claude app, writing a few Markdown skills and a little Python but mostly describing what he wanted and correcting the output. A scheduled task runs every day at 6:20 AM. It reads her dedicated job alert inbox (fed by LinkedIn, Naukri and IIMJobs alerts), an Indeed connector, two Apify actors, the Parallel Search connector and the careers pages of about 50 target companies. It filters roles against a criteria file kept in GitHub, merges duplicates across sites into one card, pulls employer ratings from AmbitionBox or Glassdoor, and scores each role against her master resume with a note on fit and gaps. Results land on a dashboard where a Tailor now button has Claude rewrite the resume and draft a cover letter. A separate Claude agent that never saw the draft checks every claim against the master resume and the PDF, and each draft and letter is versioned in its own GitHub folder.

In the first week it read a few hundred postings each morning and put 18 roles on the dashboard. She shortlisted 6, applied to 4 and skipped 5. Six tailored drafts were made and four passed the release check; a draft gets two rounds of fixes before it is held back with the problems noted. One Apify actor used over half the budget without finding any role the free sources missed, so he dropped it. She said the resume tailoring saved her the most time, followed by seeing company ratings next to each role.

It runs on his existing Claude Pro plan, which needs scheduled tasks; she uses a free account just to work the dashboard. Apify costs about $7 a month, covered so far by its free credit, and can be dropped in favour of email alerts plus a free search connector. He warns not to trust the tailored resumes or cover letters without reading them, and she still decides and applies herself. The code is published at github.com/kalitab/morning-job-scan.