Skip to content
L-03Project ArchiveEL +0.00 m

Backend & API Projects

Backend, API, payment, and developer tooling projects built with Node.js, TypeScript, Express.js, AWS, Docker, Redis, and PostgreSQL. Includes in-depth engineering case studies for Lesoll and EgyStay — production marketplace and booking backends — plus SRVJ, a real-time CRDT collaboration server. Each deep dive covers the architecture, production incidents, and engineering decisions behind it.

GitHub
Lesoll logo

Lesoll

Production B2B/B2C real-estate & automotive marketplace backend

Read the deep dive →

Technologies

Typescript Javascript Express.js MongoDB Paymob (Payment Gateway) Resend BullMQ Redis Puppeteer Swagger Docker AWS (S3, EC2, CloudFront, Route 53) Socket.IO NGINX pm2 pnpm git/github Github Actions
  • Lesoll is a large-scale classifieds marketplace for the Egyptian real estate market, enabling users to buy, sell, and rent residential, commercial, land, and compound properties.
  • The platform includes premium listing packages, real-time messaging, a blog, and a complete internal administration system.
  • As the primary Backend Engineer, I designed and implemented core backend services, REST APIs, business logic, database architecture, third-party integrations, and performance optimizations.
  • I owned 100% of the backend implementation, building and maintaining features across the listing lifecycle, authentication, payments, notifications, search, analytics, and internal admin services while supporting a production platform serving real users.
EgyStay logo

EgyStay

Property rental & booking platform with a Co-Host marketplace

Read the deep dive →

Technologies

Typescript Express.js MongoDB Amazon Payment Service (APS) Paymob Resend BullMQ Redis Swagger Docker AWS (S3, EC2, CloudFront, Route 53) Socket.IO Axios NGINX pm2 pnpm git/github Github Actions
  • EgyStay is a short-term rental platform for the Egyptian market, allowing guests to book accommodations while hosts and co-hosts manage their properties through dedicated dashboards.
  • As the primary Backend Engineer, I architected and developed the backend from the ground up, implementing reservation workflows, availability management, pricing logic, payments, authentication, messaging, notifications, and administrative services.
  • I contributed approximately 80% of the project's implementation (1100 of 1300 commits), delivering around 63,000 lines of production code across 600+ source files while focusing on scalability, maintainability, and performance.
SRVJ logo

SRVJ

Real-time collaborative diagram editor built on Yjs CRDTs

Read the deep dive →

Technologies

TypeScript Express PostgreSQL Prisma MongoDB Redis BullMQ Docker Kubernetes NGINX Socket.IO CRDTs (Yjs) Cloudinary Vitest Vue.js Vite Vue Flow Pinia UnoCSS Iconify Github Actions git/github PNPM
  • Built real-time collaborative diagram editing using Yjs CRDTs over a custom WebSocket server implementing the Yjs sync and awareness protocols, enabling live cursors, presence, and seamless multiplayer editing.
  • Secured WebSocket connections with PASETO v4 authentication and enforced project-level role-based access control, while isolating collaboration sessions into dedicated board rooms.
  • Designed a dual-snapshot persistence layer that stores both the authoritative Yjs binary state for lossless recovery and a denormalized JSON representation for efficient querying and API reads.
  • Architected the backend using PostgreSQL/Prisma for relational data (users, projects, memberships, sharing) and MongoDB/Mongoose for high-frequency diagram mutations.
  • Implemented horizontally scalable notifications using SSE with Redis Pub/Sub fan-out, alongside BullMQ workers for asynchronous email and notification processing.
  • Improved platform reliability and security through Zod validation, Helmet, CORS, rate limiting, centralized error handling, structured logging with Pino, and Prometheus metrics. Verified collaboration behavior with integration tests and containerized the platform using Docker Compose behind Nginx.
Gen Import logo

Gen Import

Automatically generate TypeScript/JavaScript barrel files using the TypeScript compiler API

Technologies

TypeScript Express pnpm git/github Github Actions fs
  • Builds a full module dependency graph (static imports, `export *`, dynamic `import()`, `require()`, class heritage, decorators) and detects circular dependencies
  • Classifies the barrel itself as safe / type-safe / ordered / unsafe, and explains exactly which line breaks it at runtime
  • Sorts files topologically so the barrel import order is safe for CommonJS
  • Optionally emits lazy (getter-based) re-exports so the barrel itself never causes a circular-require crash
  • Detects export name collisions across source files
  • Diffs against the previous barrel to report newly added exports
  • Prints diagnostics (with fix suggestions), a summary box, and an import/export graph, using `boxen` + `chalk`
KeepITs logo

KeepITs

Lightweight schedule & task management web app

Technologies

TypeScript Vue.js Vite Firebase Authentication Cloud Firestore Vercel Responsive Design Task Management Schedule Management Productivity GitHub

A lightweight schedule management web app built for organizing daily tasks and dates in a simple, clean interface. The application is deployed on Vercel, providing fast loading, automatic deployments, and reliable hosting for modern web projects.

0Gosha logo

0Gosha

Rule-based prompt-optimization engine that learns from feedback — no LLM API calls

Technologies

TypeScript Express.js MongoDB Redis BullMQ nodemailer Swagger Docker Vitest Zod cloudinary Axios NGINX pm2 Socket.IO Gen-Import pnpm git/github Github Actions
  • Pure AI agent engine that optimizes prompts and learns from user feedback.
  • No AI API calls. No OpenAI. No Claude API.
  • 0Gosha is a self-contained rule-based engine that rewrites raw user text into professional, structured prompts optimized for Claude, GPT, or any LLM.
  • It gets smarter over time by learning which transformation rules produce the highest-rated results.

Elrecord

Scalable real-time chat backend inspired by Discord

Technologies

TypeScript Express.js PostgreSQL Prisma Amazon Payment Service (APS) Message Queue (BullMQ) Swagger Redis Socket.IO Docker AWS S3 NGINX pm2 pnpm git/github

Elrecord is a powerful, scalable backend for a real-time chat application inspired by Discord. It supports direct messaging, group chats, and server-based communities with voice/video call capabilities.