Hello

@Code by Arpit Gupta

CUD

A fast, privacy-first platform for meaningful social connections.

carousel-0carousel-1carousel-2

Overview

CUD is a next-gen social discovery platform designed for authentic, intentional interaction. It breaks the pattern of shallow engagement by fostering deeper connections and real-time, meaningful conversations. The platform includes a real-time chat system, intelligent matchmaking, secure payments, and a full image moderation pipeline across multiple microservices.

Role

Founder & Full-stack Architect – led the end-to-end design and development of scalable backend services, high-performance real-time systems, and moderation infrastructure.

Key Features

  • GraphQL-powered backend with typed schema-first development
  • Secure WebSocket-based chat system (Actix + actix-ws) with reconnect support and E2E encryption
  • Dedicated microservice using Nimbus for handling Razorpay-based payments, order verification, and webhook routing
  • Image moderation pipeline using a custom Node.js service with NSFW filtering and review flow
  • Admin dashboard built in Next.js for manual moderation and flag resolution
  • Presigned media upload system with AWS S3 and Cloudflare acceleration
  • Notification service using Redis streams for async delivery of user events
  • Real-time user matching and connection suggestions based on location and personality profiles

Tech Stack

  • Rust (Actix, Tokio)
  • async-graphql
  • WebSockets (actix-ws)
  • Redis (Fred)
  • MongoDB
  • Next.js (Moderation portal)
  • Node.js (Moderation & ML inference)
  • AWS S3 + Cloudflare R2
  • EC2 (Self-managed infra)
  • Nimbus (Rust microservices framework)
  • Razorpay SDK
  • Terraform (infrastructure as code)
  • firebase (for real-time updates)

Challenges Solved

  • Architecting a performant, single-process WebSocket backend in Rust using Actix
  • Building a GraphQL API over a non-blocking async ecosystem
  • Securing media upload workflows with AWS S3 and Cloudflare while ensuring moderation compliance
  • Maintaining strict privacy and performance benchmarks across real-time features

Outcome

Successfully launched a robust MVP with ultra-low-latency chat and real-time connection suggestions. Moderation pipelines, payment integration, and utilities like notifications and analytics were fully functional and production-ready.

Learnings

Deepened my expertise in async Rust (Actix, Tokio), secure GraphQL APIs, real-time communication, and multi-service architecture. Built modular components for authentication, moderation, and notifications with clear interfaces across microservices. Learned to coordinate services efficiently using Redis streams, pub-sub, and presigned S3 workflows.

Demo