Quickly tweak your resume to for other job descriptions using AI agents!

TaiLOR – AI Resume Tailoring Tool

Inspiration

Two similar job descriptions but with a change of a few keywords or frameworks, and you have to edit the resume once again? Why not use TaiLOR, which tweaks resumes using a bunch of AI models!

What it does

TaiLOR takes an existing resume and, along with a job description you supply, tweaks the resume to add the technology stack, relevant frameworks, or polish and highlight the desired keywords of the job description your resume is missing!
With just the button, you get a polished revised resume in PDF format in the end, which you can further edit manually or submit as it is. My project provides in-browser integration for viewing the resume in both .tex and .pdf versions in real time, so one can see and edit the changes on the go!

How I built it

My tech stack is a Node.js/Express-based interface, and I used Warp AI to prototype the design and architecture.
I used OpenAI API and Martian API to route to an array of AI models, which generate the best result for my resume.
I used pdflatex for .tex to PDF compilation and integrated/connected the .tex and compiled PDF files in the backend to the frontend UI browser viewing of both .tex files and PDF files.

Challenges I ran into

I initially started with adding and authenticating Overleaf in-browser, but its time limit in compiling the LaTeX led me to pivot to local .tex compilation.
Adding the routing option for multiple agents was a hassle, but I used Martian to get the choice of handling the best models.
Some time was also spent debugging PDF compilation errors while using pdflatex.

Accomplishments I’m proud of

Solo'ed the project in 14 hours after networking, swag collection, fun workshops, and trading competitions!

What I learned

First time building an AI agent tool, and using a bunch of models to produce the perfect response.
Learned more about prompt engineering, automation, and dependency management.

What’s next for TaiLOR

An immediate to-do would be to give users the option to choose model options from Martian and tune other parameters using command line or in-browser integration!

From a technical perspective, for editing a small section of a resume, one can fragment the resume into targeted code chunks and send that to models for getting a response, saving on time and resources. Another feature I want to work on is integrating a web crawler for job search and parsing important keywords, and an automation script for auto-applying to jobs.

A one-in-all Job Application tracker!