MohammadWaleedFrdoose

The whole platform a donut chain runs on. Customers order from the app and website, the call center takes phone orders, branches sell through POS and waiter apps, factories produce, warehouses supply them, and one admin panel manages it all.

• Customer app and website: sign-up with SMS codes, menus and prices per branch, delivery and pickup, saved addresses, card payments (PayTabs, MEPS), wallet, loyalty points, gift cards, ratings and push notifications.
• Call center: phone orders for any branch, customer lookup, complaints, call records and agent reports.
• Branches: POS, waiter and branch apps with tables, shifts, Z-reports and inventory counts. An offline server keeps selling without internet, drives card terminals and kitchen printers, and syncs when the connection is back. Cashier screens are recorded and played back as video streams (HLS), and staff checklists are checked by location.
• Factories, warehouses and maintenance: branch orders to factories, recipes and production, barcodes, Arabic PDFs, purchase orders and stock ledgers, embedded Metabase dashboards and maintenance tickets, with one sign-in across every panel using single-use signed tickets.
• Admin panel: one place to manage menus and branch prices, offers, employees and roles, every order, the call center, delivery providers, WhatsApp campaigns, app content, reports and the audit log.
• A WhatsApp Business module on Meta's Cloud API: message templates, broadcasts to contact lists imported from Excel, signed webhooks for replies and delivery status, and the cost of each campaign.
• Delivery platforms (Careem, MyThings) with a menu sync that retries on failure, order intake and status updates; Jordan's e-invoicing (JoFotara) recorded in an append-only ledger; an audit log of every field changed and by whom.

My part

Backend for every part: designed and built. Admin panel (Angular): I added pages such as reports, WhatsApp broadcasts, QR sessions, table order history and the company factory report.

Stack

Java 21, Spring Boot, PostgreSQL, Redis, RabbitMQ, Flyway, Docker, GitHub Actions, AWS EC2, Cloudflare, k6, Angular

A pharma factory has to control every procedure: who wrote it, who approved it, which version is in force today. QBridge does that for many companies, each in its own database.

• A separate database per company, created automatically, with tests proving one company can never reach another's data.
• Sign-in through Keycloak with Google, Microsoft and passkeys. Approvals are signed electronically after a fresh password check and tied to the document's SHA-256 hash.
• The full life of a document: drafts, versions, check-out and check-in, review and approval rounds, change requests, periodic reviews, controlled paper copies and an append-only audit trail.
• Word templates filled in automatically, inspection PDFs with a signature page, uploads checked by their file signature, bilingual email reminders, Excel import and migration from older systems.
• Isolation and architecture rules tested with Testcontainers and ArchUnit.

My part

Backend: designed and built. Frontend: features in Angular (document library, passkeys, document types).

Stack

Spring Boot, Spring Security (OAuth2), Keycloak, PostgreSQL, Flyway, Cloudflare R2, Nginx, Docker, Angular

Everything an HR team does each month, from attendance to payslips, for many companies on one platform.

• A modular monolith on Spring Modulith, where modules talk through events instead of calling each other.
• Each company's data kept apart by PostgreSQL Row-Level Security.
• Payroll for Jordan and Saudi Arabia: GOSI and social security, end-of-service benefits, tax brackets, payslips and bank files, with Hijri dates, Ramadan hours and each country's holidays.
• Leaves with accruals and carry-over, attendance with geofence and photo, shifts and overtime, calendar feeds, approval chains, assets, appraisals, loans, expenses and HR letters.
• An AI assistant (Claude with tool use) that proposes changes to records as a diff. Nothing is saved until the user confirms, and answers stream over SSE.
• Moves companies over from ZenHR through its API: employees, salaries and leave balances, reconciled before go-live.
• Two-factor sign-in, Argon2 passwords, rate limiting, idempotency keys, Web Push, and scheduled jobs that run once across servers (Quartz, ShedLock).

My part

Backend.

Stack

Java 25, Spring Boot 4, Spring Modulith, PostgreSQL (RLS), Redis, Quartz, ShedLock, Bucket4j, Claude API, Docker

Scan a supplier invoice and the system reads it. Ask a question about spending and it answers from the data.

• AI reads invoices and cheques from photos, PDFs or straight from the office scanner into structured JSON (multimodal Gemini). The user checks the draft before it is saved.
• An AI chat with 7 tools it can call: search invoices and cheques, supplier spend, unpaid invoices, dashboards, and free questions answered with SQL over 25 read-only views, limited to what the user is allowed to see.
• Price-offer advice (buy or negotiate) from purchase history, automatic fallback to other models on rate limits, and every AI action written to the audit trail.
• Cheques, supplier payments and advances, and price alerts.
• A small desktop agent that prints receipts (ESC/POS) and scans from network and Windows scanners, shipped as a Windows installer and a Debian package.

My part

Backend and the print agent. Frontend: AI scanning pages and the audit log in Angular.

Stack

Spring Boot, PostgreSQL, Gemini API (multimodal input, function calling, JSON output), Cloudflare R2, Docker, GitHub Actions, Angular

Stock, production and customers for companies that make and sell their own products.

• A report builder that turns every database table into a report source, with grouping, totals, master-detail and Excel or CSV export, labelled in Arabic.
• A multi-tenant warehouse: a database per company, an admin console, and a Cloudflare DNS record created for each new company's subdomain.
• A CRM for call centers: customers, loyalty points, wallets, complaints, reviews and Excel import.

My part

Backend. Frontend: the tenant console and tenant-aware sign-in in Angular.

Stack

Spring Boot, PostgreSQL, Flyway, Apache POI, Metabase, Cloudflare API, Nginx, Docker, Angular

A marketplace where stores and restaurants post offers and deliver orders.

• 10+ microservices behind an API gateway with service discovery.
• Stores, menus, a Redis cart, orders, drivers, taxes and discounts.
• Nearby stores found with MongoDB geospatial queries; delivery cost and time from Google Distance Matrix and Directions; delivery areas and blocked zones.
• Customer and store chat on Firebase Realtime Database, live call updates over SSE, push notifications and SMS sign-in.

My part

Backend.

Stack

Spring Cloud (Gateway, Eureka, OpenFeign), MongoDB, Redis, RabbitMQ, Firebase, Google Maps APIs, Docker

Restaurants and shops subscribe and get their own online store.

• Each customer gets its own subdomain, theme and logo, routed by Nginx.
• Catalog, options, offers, cart, wishlist, discounts, loyalty points, orders, delivery integration, WebP images and video.
• The customer is carried through every service call, so each request only sees its own store's data.
• SMS and email providers set per customer.

My part

Backend.

Stack

Spring Cloud, MongoDB, PostgreSQL, Redis, Nginx, Docker, GitHub Actions

Parents add their children and the app keeps track of every dose.

• Schedules each child's doses and reminds parents by push notification, SMS and email.
• Paid subscriptions through PayTabs with promo codes, and Excel reports for admins.

My part

Backend.

Stack

Spring Boot, MongoDB, Firebase, PayTabs, Docker