easyATS

Wonder why your resume’s been ghosted? It’s probably chilling in an ATS dumpster. Our FAANG-simulated ATS (powered by Claude) scans, roasts, and revives it, before recruiters click “nope.

Comment

Inspiration

What it does

easyATS simulates FAANG-style Applicant Tracking Systems using Claude AI to scan your resume, roast it with brutal honesty, and give tailored, company-specific feedback. Oh, and it’ll even write a custom cover letter while it’s at it.

How we built it

We used Streamlit for the frontend, PyMuPDF to parse uploaded PDF resumes, regex magic to match keywords and quantify impact, and Anthropic’s Claude API to simulate a smart, sarcastic ATS reviewer. We built custom keyword+focus profiles for Meta, Amazon, Apple, Netflix, and Google. Cover letters are generated on-demand in Claude’s finest HR-speak.

Challenges we ran into

  • Claude was too nice at first—getting it to roast resumes like a FAANG recruiter took some prompt engineering finesse.
  • Some PDFs came out garbled, so we had to improve text extraction and formatting.
  • Rate limits are real. We learned to cache and batch like pros.

Accomplishments that we're proud of

  • Claude’s feedback is actually useful (and kinda funny).
  • Our scoring system genuinely reflects what tech recruiters look for.
  • Users walk away with an actual game plan to improve their resumes—plus a bonus cover letter.

What we learned

  • ATS systems aren’t evil—they’re just picky.
  • Prompt engineering is everything when working with LLMs.
  • People really do love watching their resumes get roasted.

What's next for easyATS

  • Add startup and nonprofit-specific ATS modes.
  • Let users upload job descriptions and tailor their resume to that exact posting.
  • Add analytics to track keyword trends across top tech companies.
  • One-click LinkedIn roast? Maybe.

Built With

  • claude
  • pymupdf
  • python
  • regex
  • streamlit
Share this project:

Updates