Roast My Resume

Tired of resume silence? Let our AI roast your resume with tough-love humor, then give you a "Glow Up Plan" with actionable advice to finally stand out and land the job.

Comment 1

🔥 Roast My Resume: From Boring Interview Prep to Resume Roaster

The Spark

It started with doom-scrolling through LinkedIn and Reddit this year. Post after post of talented people losing their jobs, struggling with interviews, asking the same question: "Why won't anyone hire me?"

I wanted to help, so I began building "Why Hire Me" - a straightforward AI tool to help people answer that dreaded interview question. Using Google's ADK framework, I created a multi-agent system that would analyze resumes and job descriptions to generate perfect interview pitches.

It worked. It was useful. It was... boring as hell.

The Pivot

That's when it hit me - job hunting is already depressing enough. What if instead of another serious career tool, I created something that would make people laugh while still helping them improve?

Enter Roast My Resume - the brutally honest resume reviewer that roasts you first, then helps you glow up.

Building with Fire

The technical journey taught me valuable lessons about multi-agent AI systems:

roastmyresume/
├── 🧠 Root Agent (The Orchestrator)
├── 🔍 Resume Analyzer (Finds the roastable content)
├── 📋 Job Analyzer (Understands what they actually want)  
├── 🔥 Roast Master (Delivers the burns)
└── ✨ Improvement Advisor (The friend who helps after)

Challenge #1: Learning to "Vibe Code" with AI

Gone are the days of solo debugging with Stack Overflow. This project taught me how to code alongside AI - "Vibe Coding." Instead of fighting through errors alone, I learned to:

  • Use AI to help debug mysterious 500 errors
  • Let AI explain ADK's quirks while I built
  • Iterate faster by having AI generate boilerplate while I focused on logic

The personal breakthrough? It's learning by making plenty of mistakes.

Challenge #2: Making AI Funny

Teaching an AI to be genuinely funny while staying helpful was like teaching a robot to dance - technically possible, but requires finesse. The solution? Five roast levels, from "Gentle" to "No Filter mode":

  • Level 1: "Your experience is... interesting"
  • Level 5: "THIS RESUME IS SO BAD IT'S GIVING ME ANXIETY!"

Challenge #3: The Scraping Dream That Died

Originally, I wanted users to just paste a LinkedIn URL. Reality check: scraping LinkedIn is like trying to break into Fort Knox with a spoon. Pivoted to copy-paste, which turned out better for privacy anyway.

Challenge #4: Parsing AI Responses

Getting four different AI agents to output in a consistent format was like herding cats. Solution: Over-engineer the parsing on the frontend because the backend was being creative with formatting.

What I Learned

  1. Humor + Help = Engagement - People are more likely to improve when they're laughing than crying
  2. Multi-agent systems are powerful - Each agent focusing on one task created better results than one mega-prompt
  3. User experience trumps features - Copy-paste beat my fancy scraping dreams
  4. Debugging distributed AI is an art - When agent A talks to agent B, C, and D, logs become your best friend
  5. ADK is incredible once you understand it - But that learning curve is steep

Final Thoughts

Sometimes the best projects come from taking something serious and asking, "What if we made this fun?"

Roast My Resume proves that career tools don't have to be boring. They can be brutal, hilarious, and helpful all at once.

Because in 2025's tough job market, we all need a good laugh before we level up. 🔥

P.S. The irony isn't lost on me that I built a resume roaster while procrastinating on updating my own resume.

Share this project:

Updates