Hireza: AI candidate matching that won CipherThon

Hireza was a hackathon project I built with a team of four in early 2024, and we won first place at CipherThon 2.0 with it. It’s an AI-powered hiring platform — at a time when the AI boom was just getting going, an AI candidate-matching tool still felt fairly novel.

What it does

The idea was to take the manual grind out of first-pass hiring: match candidates to a job role, and analyze resumes for how well they fit (including an ATS-style check). We used OpenAI to power the matching and resume understanding.

Who did what

I want to be accurate about scope, since it was a team effort. Two of us did the bulk of the development. One teammate came from an AI/ML background and led the OpenAI integration; I built the frontend and backend (MERN stack) and worked with him to wire the AI in — the candidate-job matching, the resume analyzer, and the interface around it.

What I took away

Hireza was an early lesson in building something AI-powered under real time pressure, and in splitting work cleanly with a teammate so the ML and the application meet in the middle. Winning CipherThon was the fun part; the lasting part was getting comfortable putting an LLM behind a real product flow, which is most of what I do now.

A note for honesty: some specific performance figures floated around for this project in old write-ups. They weren’t real measurements, so I’ve left them out here — the project stands on what it did and the win, not invented numbers.