Resume Craft

Cursor for LaTex resumes Vibecode your LaTex Resumes now

Comment

Inspiration

This idea was born out of pure frustration.

While I was applying for jobs, I realized I was spending 15–20 minutes tailoring my resume for every single position. Doing it manually in LaTeX was painful — editing, reformatting, recompiling — over and over again.

Then I tried using AI. It helped, but it turned into a long loop:
Ask Gemini to modify the code → copy to Overleaf → compile → fix errors → repeat.
What should’ve taken two minutes was taking ten.

That’s when it clicked — why can’t we have something like Cursor, but for LaTeX resumes?
An editor where AI understands what you want, edits your LaTeX code, fixes the syntax, and compiles it instantly — all in one place.

That frustration turned into Resume Craft.

What it does

Resume Craft is an AI-powered, LaTeX-integrated resume builder.

It lets you chat with AI in plain English to update, rewrite, or tailor your resume for any job — instantly.
You tell it what you want (“make my resume sound more data-science focused” or “add my latest project”), and it updates your LaTeX code, fixes errors, and compiles your resume automatically.

No more jumping between Overleaf, editors, or debugging LaTeX errors.
It’s like having your personal AI co-pilot for resume editing.

How we built it

I used Gemini 2.5 Flash API for real-time LaTeX code editing and optimization.
The backend is built with Python, hosted on Google Cloud Run for a serverless, scalable setup.
The frontend is built using React, TypeScript, HTML, and CSS, and hosted on Netlify.
I also integrated Razorpay for future premium upgrades or donations.

All of this together forms a full-stack AI-driven compiler — powered by Gemini, built to save people time.

Challenges we ran into

Making AI-generated LaTeX compile flawlessly was a serious challenge.
AI models often output small syntax errors that completely break the PDF build. I had to design a feedback loop where Gemini analyzes and fixes its own compilation errors — without user intervention.

Another challenge was latency and deployment — handling real-time responses from Gemini on Google Cloud while keeping the compiler snappy and responsive.

And of course, building a clean UI that doesn’t overwhelm non-technical users took quite a few redesigns.

Accomplishments that we're proud of

The biggest win was seeing the entire loop come to life — typing in natural language and instantly getting a valid compiled resume PDF.

I’m proud that Resume Craft has already started gaining real users and positive feedback.
Seeing people use something that started as a small personal pain point is honestly the best feeling.

Building a working AI compiler platform — solo — was a huge milestone for me.

What we learned

I learned that building with AI isn’t just about prompts — it’s about product design and UX clarity.
You need to make complex systems feel simple.

I also learned a lot about prompt control for code generation, error recovery, and handling compiler edge cases gracefully.
Most importantly, I learned that real innovation often starts with fixing your own problem.

What's next for Resume Craft

Next, I want to:

  • Add AI model selection, so users can switch between Gemini, Gemini Flash, GPT, or Claude.
  • Introduce a share option, letting users send or publish their resumes easily.
  • Build a feedback system so the AI learns from user edits and gets smarter over time.
  • Expand access and launch Resume Craft publicly to help more job seekers save hours and land interviews faster.

What started as my personal workaround is now turning into a tool that helps others vibe-code their dream resumes.

Share this project:

Updates