This is a submission for Weekend Challenge: Passion Edition
What I Built
My biggest passion is the developer community and hackathons. I love seeing students build amazing things, but Iβve also seen the struggles: organizers stress over check-ins, and beginners freeze when they can't think of a good project idea.
Fueled by this passion to help my community, I built HackDesk - Passion Edition. It is an intelligent, high-performance event management system designed for large-scale hackathons. It features a robust attendee registration system and an AI-powered Hackathon Idea Generator that instantly provides customized, beginner-friendly project ideas to students so they can start building immediately.
Demo
Live Vercel Deployment: HackDesk Passion Edition
(Feel free to try the Idea Generator to see the Google AI fallback in action!)
Code
π HackDesk - Passion Edition
An Intelligent, Concurrency-Proof Hackathon Management Platform
HackDesk is a high-performance event management system designed to seamlessly handle large-scale hackathons. It features a robust registration and check-in system, alongside a unique AI-powered Hackathon Idea Generator equipped with atomic locking to prevent race conditions during high-traffic claim events.
Built with modern web technologies to ensure low latency, high availability, and a premium user experience.
β¨ Key Features
π₯ Smart Registration & Attendance: Bulk-import participants via CSV parsing Manually add late or on-the-spot attendees seamlessly. Active Session tracking (e.g., "Day 1", "Day 2") for accurate multi-day check-ins.
π§ AI-Powered Idea Pool (Pre-generation & Smart Fallback Architecture): Organizers can generate batches of beginner-friendly project ideas based on specific themes using the Google Gemini AI API Pre-generating ideas prevents API rate limits and keeps the end-user experience lightning fast. If the pool runs empty, the system uses Google AI asβ¦
How I Built It
I built HackDesk using Next.js 16 (App Router), TypeScript, Tailwind CSS, and Supabase (PostgreSQL).
Managing a hackathon means dealing with high traffic (e.g., 200 students clicking "Generate Idea" at the exact same millisecond). To solve this, I engineered a Concurrency-Proof Idea Claiming system. The backend uses strict isolation and atomic row-level locking in PostgreSQL so that no two students ever get the same idea.
The Smart Fallback Architecture:
To ensure the system never crashes or runs out of ideas, I implemented a two-tier architecture:
- It first tries to fetch a pre-generated idea from the Supabase pool.
- If the pool is completely empty, instead of throwing an error, the system seamlessly falls back to Google Gemini AI 1.5 Flash. The AI generates a fresh, structured JSON idea on the spot, saves it to the database, and assigns it to the student instantly.
Prize Categories
I am submitting this project for the Best use of Google AI prize category.
I utilized the @google/generative-ai SDK (Gemini 1.5 Flash model) to create the live fallback mechanism. This ensures that the hackathon's idea generation never hits a dead end, combining database reliability with the dynamic generative power of Google AI.
Top comments (0)