CoopSync Assistant

Inspiration

Students often spend hours manually comparing their resumes with hundreds of job postings, so we wanted to simplify this process with AI.

What it does

CoopSync Assistant analyzes job listings and compares them with a user’s resume to generate a match score and a short explanation.## How we built it We built a Chrome extension that extracts job data from the page and sends it with the resume to an AI model through the OpenRouter API.

Challenges we ran into

Our biggest challenge was a broken data flow: the dashboard was reading only local jobs.json, so live WaterlooWorks-exported data was not actually used for AI scoring. The second major bug was repeated re-scanning in the content script: opening the side widget could trigger another analysis cycle, causing duplicate requests and unstable UX.

Accomplishments that we're proud of

We created a working AI tool that integrates directly into a job board and automatically evaluates job-resume compatibility.

What we learned

We learned how to build Chrome extensions, design AI prompts, and manage asynchronous communication with external APIs.

What's next for CoopSync Assistant

We plan to support more job platforms and add features like resume improvement suggestions and personalized job recommendations.

Built With

Submitted to

EngHacks
- Winner First Place - Per Person On Team

Created by

-Built and stabilized handleScoreJobsForDashboard to score each job against resume data.
-Added sanitizeDashboardJob to normalize job inputs before scoring.
Added summarizeHiringHistory to compress hiring-history context into prompt-friendly text.
Integrated callOpenRouter + getStoredProfile to securely fetch credentials/resume and call the external AI API.
Enforced reliable output shaping: score clamping (0–100), reason fallback, and consistent { posting_id, score, reason } responses.

Sissie Qu
I wrote scripts to scrape detailed job information and data from WaterlooWorks, including publicly visible data such as job title, company name, and number of applicants, as well as hidden data such as work term rating data, detailed job descriptions, and requirements. I organized and structured this data so that it could be exported or passed to other parts of the project.

I also developed a script that allows a specific WaterlooWorks job posting page to be opened externally from anywhere using the job ID. In addition, I participated in most of the program integration work.

Yuyang Mei
I worked on the front-end dashboard and the communication logic between the dashboard and the WaterlooWorks page. It was my first time organizing a project into multiple helper functions and building a modular structure in JavaScript. At first it was a little challenging to manage the different modules and browser extension APIs, but through this project I learned how to structure code more clearly and build a more maintainable system.

Zhimo2120

Updates

Sissie Qu started this project — Mar 07, 2026 11:03 PM EST

Leave feedback in the comments!

Log in or sign up for Devpost to join the conversation.