AI-Driven Race Photography Platform Development

This project involved the end-to-end design, development and deployment of RunThrough Photos, an enterprise-grade race photography platform. The system cloud-based AI to solve the primary challenge for race participants- finding their photos quickly and accurately within a gallery of thousands of images. Built on a modern, full-stack Next.js 14 architecture, the platform features a public-facing gallery with intelligent search capabilities and a comprehensive administrative dashboard for event, user, and analytics management

Frontend- Next.js 14 (App Router), React 18, TypeScript, Tailwind CSS, Radix UI

Backend- Next.js API Routes, NextAuth.js, Socket.IO, Server-Sent Events (SSE)

Database & BaaS- Supabase (PostgreSQL, Realtime, RLS)

Cloud & AI Services- AWS S3, AWS Rekognition, AWS Textract

Image Processing- Sharp

Deployment- Vercel

1. The core of the platform is its ability to search for photos using advanced AI, providing users with a fast and intuitive discovery experience

Utilizes AWS Rekognition to index and search for faces. Users can upload a selfie to instantly find all their photos in an event gallery, with results ranked by a confidence score

Employs AWS Textract to perform Optical Character Recognition on images, detecting and indexing race numbers with high accuracy

2. A fast, responsive, and feature-rich gallery system designed to handle thousands of images efficiently

Photos are logically organized by month and event, with dynamic routing for clean, shareable URLs

Features a responsive photo grid, pagination for efficient loading, and an optimized full-screen modal viewer

A server-side system using Sharp applies custom, event-specific watermarks to photos before download, with configurable position and sizing

3. A secure, role-based administrative area for complete control over the platform's content and operations

Provides live statistics on user registrations, photo downloads, and search query performance

Allows admins to manage events, monitor photo counts, and manually trigger or monitor the AI indexing processes for faces and race numbers

Full capabilities for managing admin and member accounts with role-based permissions, including event-specific access delegation

4. The platform is built with scalability, security, and performance at its core, incorporating enterprise-level patterns

An S3 webhook triggers an asynchronous background process to automatically analyze and index new photo uploads, ensuring galleries are always up-to-date

Integrates Socket.IO and Supabase Realtime to push live updates to both the admin dashboard and public galleries, reflecting indexing progress or new photo additions without requiring a page refresh

Implements database-level Row Level Security (RLS) in Supabase, secure JWT-based authentication, and extensive performance optimizations, including image optimization (WebP/AVIF), code splitting, and database indexing