JudgeJam – AI-Powered Hackathon Judging Platform

Overview

JudgeJam is an AI-assisted hackathon judging platform designed to streamline the evaluation process for large-scale hackathons. Traditional judging requires significant manual coordination and can be time consuming for organizers and judges. JudgeJam automates parts of this workflow by collecting project submissions, analyzing presentation content, and generating structured evaluations.

The platform uses machine learning and natural language processing to help organizers process submissions more efficiently while still supporting human decision-making.This project was built to explore how AI systems can assist with decision workflows in event management environments such as hackathons.

Features

Real-Time Pitch Analysis

• Analyzes hackathon pitch presentations using computer vision and audio processing to evaluate facial expressions, gestures, tone, and filler words.

AI-Powered Project Evaluation

• Processes project descriptions, presentations, and repository data to generate structured scoring and feedback.

Pitch–Project Consistency Checking

• Uses language processing to detect inconsistencies between a team's pitch and their submitted project materials, helping judges evaluate clarity and authenticity.

Automated Judging Workflow

• Aggregates submissions, presentations, and evaluation metrics into a single system, reducing manual review time for judges.

GitHub Repository Verification

• Integrates with the GitHub API to verify repository creation and activity timestamps, ensuring projects comply with hackathon rules.

Centralized Project Dashboard

• Compiles all project materials, evaluations, and metadata in one place to streamline the judging process.

Tech Stack

Frontend

• Next.js
• JavaScript

Backend

• Node.js
• Express
• MongoDB

AI / Machine Learning

• PyTorch
• OpenCV
• DeepFace
• MediaPipe

APIs & Cloud

• OpenAI API
• GitHub API
• Google Cloud