've been building in the job search space for a while now. But before I built anything, I went through the job hunt myself — and like most people, I did it wrong.
I had one resume. A good one, I thought. Clean formatting, strong bullet points, quantified achievements. I applied to 40 jobs over the course of a month. I got four callbacks. One offer.
That's a 10% callback rate, which is actually around average. But here's the thing — it shouldn't be average. A 90% rejection rate, most of it before a human reads your resume, isn't a "numbers game." It's a systems problem.
Here's what I learned after going deep on how ATS works:
Every job description is a scoring rubric. The ATS isn't reading your resume to understand you — it's comparing your text to the job description's text. The more overlap in specific phrases, the higher your score. "Revenue growth" and "grew revenue" are not the same to a parser. "Led a team" and "managed direct reports" are not the same. The exact language matters more than the substance behind it.
Most people are writing one resume for many jobs. That resume is written in your language, not the language of each specific role. So for each application, you're submitting a document that isn't tuned to the scoring rubric. Every. Single. Time.
The fix is tailoring — but tailoring manually is brutal. To do it right, you'd read the job description carefully, identify 15–20 critical keywords and phrases, cross-reference your experience, rewrite your bullets to naturally incorporate those phrases, and repeat for every application. That's 30–60 minutes per job. For 40 jobs, that's up to 40 hours of work on top of everything else.
Nobody actually does this at scale. So most people just apply with a generic resume and accept the low callback rate as normal.
I built Resuma (resuma.sarthum.com) because I wanted to make the "right" approach actually viable. You paste a job description, upload your resume, and it generates a tailored version — matched keywords, rewritten bullet points that mirror the job description's language, ATS-compatible formatting. It also scores your current resume against the role so you can see exactly where you're losing points before you apply.
It's not magic. It's just doing the systematic work that most people skip because it's too time-consuming.
There's also a LinkedIn optimizer, skills gap analysis, interview prep, and a video pitch analyzer built in — because resumes are only one piece of the job search, and the same keyword-gap problem shows up everywhere.
It's free to start — no credit card. We've had 1,250+ reviews with a 4.9/5 average, which honestly tells me the problem is as real as I thought it was.
One stat I didn't expect: the users who see the biggest improvement aren't the ones with weak resumes. They're the ones with strong experience who were just submitting it in the wrong format or language. The underlying work was solid. The presentation just wasn't tuned for how the system actually evaluates it.
If you're building in the career tools space — what patterns have you noticed about where job seekers actually lose ground? And if you've been through a recent job search, did you tailor for every application or go with a single version?
→ resuma.sarthum.com
INDIEHACKERS POST
Section: Products > Resuma > New Update
Title: 1,250+ reviews and 4.9/5 stars: what I've learned building an AI resume tool people actually use
What Resuma does
Resuma is an AI-powered ATS resume builder. You upload your existing resume and paste a job description, and it generates a tailored version — matched keywords, rewritten bullet points in the language of the job, ATS-compatible formatting. It also scores your resume against the role, shows you where you're losing points, and tells you how to fix it. There's a LinkedIn optimizer, skills gap analysis, interview prep, and a video pitch analyzer all under one roof.
It's live at resuma.sarthum.com. Free to start, no credit card required.
Why I built it
The problem is pretty simple once you see it: over 90% of companies run resumes through ATS software before any human sees them. These systems score your resume against the job description. If you submit the same resume for every job — which is what most people do — you're almost guaranteed to score poorly on most applications. Not because your experience is weak, but because the language doesn't match.
Fixing this manually takes 30–60 minutes per application. Nobody does it consistently at scale. So I built something that does the systematic work automatically.
Top comments (0)