← Back to Roast My Resume
Experiment

I Made AI Roast 50 Resumes. The Average Grade Was a D. Here's Why Recruiters Are Deleting Yours in 6 Seconds.

May 21, 2026 · 8 min read · TinyTools

Recruiters spend an average of 6 to 7 seconds scanning a resume before deciding to keep it or delete it. I've known that stat for years. I thought it applied to other people's resumes.

Then I ran my own resume through an AI roast tool. It gave me a D+. Said my bullet points read like a job description, not an accomplishment list. Called my summary "aggressively forgettable." Found three buzzwords in the first four lines.

I was mildly offended. Then I ran 49 more resumes — collected from friends, Reddit threads, career forums, and anonymous submissions — to see if mine was an outlier or the norm.

It was not an outlier.

78%
of 50 resumes tested scored D or below. Only 4 resumes earned a B or higher.

Here's the full breakdown of what the AI found — and why the same five mistakes showed up across nearly every single resume, from entry-level to senior director.

The Grades: What 50 Resumes Actually Looked Like

I ran each resume through the Roast My Resume tool, which evaluates resumes across impact language, specificity, ATS compatibility, structure, and keyword optimization. Here's how the 50 resumes distributed:

Grade Count % of total Common pattern
A 1 2% Quantified every bullet, clean structure, tailored keywords
B 3 6% Good impact language, minor formatting issues
C 7 14% Some metrics, inconsistent tense, weak summary
D 28 56% Duty-focused bullets, no metrics, generic summary
F 11 22% Buzzword soup, responsibility lists, ATS-unfriendly layout

The single A-grade resume came from a product manager who had quantified every single bullet point — not with vague "improved efficiency by X%" but with specific, verifiable claims like "reduced onboarding time from 14 days to 6 days for a 200-person engineering team." The roast had almost nothing to say. That's the target.

The 5 Mistakes That Killed 78% of Resumes

Across all 39 D-and-F resumes, five patterns appeared with near-total consistency. Not "most of them." Almost every single one.

1. Responsibility Bullets Instead of Accomplishment Bullets

The most common failure by a significant margin. 43 of 50 resumes led with what the person was responsible for, not what they actually achieved.

🔥 Actual roast output

"'Responsible for managing social media accounts' — congratulations, you've described a job posting, not a career. What happened to the accounts? Did followers go up? Down? Did they spontaneously combust? Give me a number, a name, or a date."

✓ The fix

Swap "Responsible for X" → "Did X, resulting in [metric or outcome]." Even an estimate is better than nothing: "Reduced customer email response time from ~48h to ~6h."

2. The Generic Professional Summary

37 of 50 resumes opened with a summary that could have applied to literally anyone in their field. The roast tool was particularly brutal here.

🔥 Actual roast output

"'Results-driven professional with a passion for leveraging synergies to drive cross-functional growth.' I've read this exact sentence 4,000 times and I still have no idea who you are or why I should care. Delete this. Start over. Who are you, specifically?"

✓ The fix

Replace it with one sentence that says: what you do, what you're specifically good at, and one number. "Backend engineer with 6 years in fintech, specialized in payment infrastructure — shipped 3 systems processing $2B+ annually."

3. Buzzword Density in the Skills Section

29 resumes had skills sections that read like a LinkedIn keyword list — "Agile, Scrum, Leadership, Communication, Stakeholder Management, Strategic Thinking." The roast called these "content-free signals."

"Strategic Thinking" appearing as a skill on a resume is the resume equivalent of listing 'breathing' as a hobby. Of course you think strategically. That's implied by being a sentient adult applying for a professional role.

The real problem: ATS systems do parse skills sections, but they're looking for specific technical skills, tools, and certifications — not personality traits. Buzzwords waste keyword space on terms recruiters don't search for.

4. Inconsistent Tense and Vague Time Periods

Present tense for past roles. Past tense mixing with present within the same bullet. Dates listed as "2019 – Present" for a role that clearly ended. These showed up in 34 of 50 resumes and signal carelessness more than anything else.

✓ The fix

Simple rule: current role gets present tense. Every previous role gets past tense. No exceptions. Check every bullet before you submit.

5. The "Duties Wall" — Bullet Points With No Structure

22 resumes had experience sections that were effectively unformatted paragraphs broken into bullets. 8-12 bullets per role, all roughly the same length, none starting with a strong action verb, no visual hierarchy. The roast called this the "I am describing my calendar" problem.

🔥 Actual roast output

"Six bullets, all starting with 'Worked with,' 'Helped,' or 'Supported.' These are supporting-character verbs. You're the protagonist of your own resume. Act like it."

The One Category That Consistently Scored Highest

Engineers who had shipped open-source projects with GitHub stars, measurable contributor numbers, or production usage metrics consistently scored highest — regardless of their job history. The AI rewarded specificity over title, every single time.

The resumes that scored B or higher shared exactly one trait: every claim was attached to a number, a name, or a date. Not a vague "improved" or "grew" — a specific, falsifiable claim.

You don't need a prestigious employer to have a high-scoring resume. You need precision.

What The Roast Actually Looks Like in Practice

The tool isn't just a score. It gives you a letter grade, flags specific red-flag phrases by category (impact language, buzzwords, ATS compatibility, structure), and delivers the feedback in a voice that's closer to a brutally honest mentor than a polite career counselor.

Some of the lines from this experiment's roast sessions were genuinely memorable. One resume earned: "This reads like it was written by someone who was very careful not to claim responsibility for anything. ATS will skip it and so will I." Another got: "Your contact section has a LinkedIn URL that goes to a 404. You've committed resume murder before the recruiter even started reading."

The savage framing is intentional. People remember feedback that stings slightly more than feedback that's polite. And the underlying analysis is real — the tool checks the same signals a trained recruiter or career coach would flag.

Try It on Your Own Resume (10 Minutes or Less)

The tool runs entirely in your browser — no upload, no account, no waiting. Paste your resume text, get the grade and the roast. The feedback is specific enough to act on immediately: you'll know exactly which bullets to rewrite, which section to cut, and which keywords are missing.

Based on this experiment: if you haven't had your resume roasted recently, statistically there's a 78% chance it's scoring D or below. That's a fixable problem — but only if you know about it.

Get Your Resume Roasted (Free)

Paste your resume text. Get a letter grade, savage feedback, and specific fixes. No signup, runs in your browser.

🔥 Roast My Resume

If you're also looking to optimize for ATS keywords specifically, the Resume Tailorer runs a different kind of analysis — matching your resume against a specific job description to find keyword gaps. And if you've been applying but not hearing back, the GitHub Roast checks whether your public profile is creating the wrong impression before a recruiter even opens your resume.