Open to Full-time & contract engineering roles

Software & AI engineer building production systems end to end.

I build across the whole stack — React and .NET, SQL and APIs — and I build AI products: agents, chatbots, workflow automation and advanced retrieval (RAG) systems. Shipping production software end to end for the last 3.5+ years.

Download CV GitHub Email

3.5+: Years shipping production code
Full-stack: React · .NET · SQL
AI: Agents · RAG · LLM apps
End to end: Design through deploy

01About

The short version.

Software & AI engineer with 3.5+ years building production systems end to end — from React interfaces through C#/.NET and SQL to AI agents, chatbots and retrieval (RAG) services. Equally at home shipping enterprise features and building LLM-powered products.

I like the parts most people avoid: untangling a legacy module, drawing a clean boundary between a command and a query, or wiring an LLM into something that has to work in production rather than a demo. I work across the whole stack — and increasingly the AI stack — because shipping a feature usually means owning all of it.

Based in: New Cairo, Egypt
Experience: 3.5+ years, production
Open to: Full-time & contract
Focus: Software & AI engineering

02Experience

Where I've shipped.

Full-Stack Software Engineer
Oct 2022 – Present
Misr Technology Services · New Cairo, Egypt
nafeza.gov.eg — Egypt's National Trade Platform
Contributing to the migration of a national-scale trade platform from legacy ASP.NET Framework to .NET 9, React 18 and Clean Architecture, inside an enterprise microservices codebase.
- Deliver features end to end across the stack: React 18 + MUI UI (multi-step wizards, forms, data tables) → DTOs → commands/queries → CQRS handlers.
- Work within an enterprise codebase using RabbitMQ, Redis, OpenIddict (OAuth2 / OIDC) and AWS Secrets Manager.
- Migrated and re-architected legacy ASP.NET Framework modules into the new .NET 9 Clean Architecture system.
- Previously maintained the large-scale legacy platform — optimized SQL queries and resolved production incidents on a critical national system.
- Co-built a custom Selenium-based regression suite automating 30–40 test cases with screenshot capture.
.NET 9C#React 18MUISQL ServerRabbitMQRedisOpenIddictClean ArchitectureCQRS
Full-Stack Engineer
Jan 2022 – Aug 2022
Self-Employed · Remote
Production full-stack web applications
Co-architected and shipped several live full-stack applications end to end — planning, data modelling, integrations, deployment.
React 18TypeScriptNode.jsExpressMySQLTypeORMTailwind CSSDocker

03Flagship

National scale, in production.

The platform I work on daily — the clearest signal of the kind of systems and stakes I'm comfortable operating at.

nafeza.gov.eg

Egypt's national single-window trade platform

nafeza.gov.eg

National scale
10+ features shipped
Microservices + CQRS

Problem

A critical national trade system running on legacy ASP.NET Framework needed to be modernized without disrupting live, high-stakes operations.

Architecture

Incremental migration to .NET 9 with Clean Architecture and CQRS, split into microservices communicating over RabbitMQ, with Redis caching, OpenIddict (OAuth2/OIDC) auth and AWS Secrets Manager. React 18 + MUI front end built feature-by-feature against the new command/query handlers.

Outcome

Delivered 10+ features end to end and migrated legacy modules into the new architecture, while a Selenium regression suite guarded against regressions across the platform.

.NET 9C#React 18CQRSRabbitMQRedisSQL Server

04AI engineering

Engineering AI products, end to end.

Alongside the application work, I design and ship AI systems end to end — agents, chatbots, automation and retrieval — with the same architecture discipline as the rest of the stack.

AI agents
Chatbots & assistants
LLM integration
Advanced RAG systems
Workflow automation (n8n)
Multilingual (AR / EN)

RAG Service

A reusable retrieval microservice for LLM apps

Repo

Advanced RAG pipeline
Pluggable providers
Multilingual AR / EN

Problem

LLMs don't know your private data, and you can't paste whole documents into a prompt. Every product I built kept needing its own retrieval layer.

Approach

A standalone Python + FastAPI + pgvector microservice built on a strategy/factory architecture — the embedder, reranker, query rewriter and vector DB are all swappable from environment variables with zero code changes. Two strategies ship: naive, and an advanced pipeline that rewrites the query → embeds → vector-searches → reranks. Tuned multilingual (Arabic + English) with Cohere embeddings/rerank and a Groq-hosted Qwen rewriter.

Outcome

One HTTP service that returns ranked context (not answers), so many products can share it — with namespaced storage, content-hash de-duplication, API-key auth, graceful provider fallbacks and a self-validating startup that refuses to run on a dimension mismatch.

PythonFastAPIpgvectorPostgreSQLCohereGroqDocker

AI Hub

Multi-channel AI chatbots & assistants

Repo

Shared engine, 5 apps
WhatsApp + web
Pluggable LLMs

Problem

Different businesses — a clinic, a supermarket — wanted conversational AI on the channels their customers already use: WhatsApp and the web.

Approach

A TypeScript pnpm monorepo with a shared conversational engine (bot-core) and a thin LLM client (ai-core) powering several apps: WhatsApp bots over Baileys and Twilio, an embeddable React chat widget, and a standalone chat UI. The LLM is pluggable — Groq and Google Gemini behind one interface.

Outcome

A reusable engine that drops a domain-specific assistant onto WhatsApp or any website, so spinning up a new vertical is a thin app over the shared core packages.

TypeScriptNode.jsGroqGeminiBaileysTwilioReact

05Selected work

Products, shipped and live.

Full builds I designed and delivered end to end — re-told for engineers: the problem, the architecture, and what shipped.

dorosna.online
01
Dorosna
Egypt's trusted digital classroom
- Teacher & student portals
- Multi-curriculum
- Bilingual EN/AR
Problem
Teachers and students across multiple curricula needed one platform to deliver and take courses online, each side with its own experience.
Architecture
A React + TypeScript front end over a Node.js backend, with role-based portals for teachers and students, a multi-curriculum course catalogue and enrolment. Bilingual EN/AR with full RTL, built mobile-first.
Outcome
A live e-learning platform connecting educators and learners across National, IGCSE, SAT and Canadian curricula.
ReactTypeScriptNode.jsTailwind CSS
Visit dorosna.online
21reve.com
02
21Reve
Furniture & interior-design e-commerce
- 39-entity model
- 57 REST endpoints
- Paymob + webhooks
Problem
A premium furniture brand needed a full storefront with real payments, custom-furniture quotations and a designer-consultation step — built end to end.
Architecture
React 18 + TypeScript front end over a Node/Express API with MySQL. A 39-entity relational model backs 57 REST endpoints; Paymob is integrated with webhook handling for live payment flows. Customer accounts, wishlist and reviews. Bilingual EN/AR with full RTL.
Outcome
A live storefront processing real orders and payments, plus a custom-quotation and consultation flow for bespoke furniture.
React 18TypeScriptNode.jsExpressMySQL
Visit 21reve.com
agnadeen.shop
03
Agnadeen
Multi-role grocery e-commerce
- Retail + wholesale
- Inventory + alerts
- Dockerized deploy
Problem
One storefront had to serve retail and wholesale customers at different pricing tiers, with an admin dashboard to run it — and ship as a reproducible deployment.
Architecture
Role-based access over a Node/Express + MySQL backend, with multi-tier pricing, inventory management and low-stock alerts. Fully bilingual EN/AR with RTL, packaged and shipped as a Dockerized deployment.
Outcome
A live grocery marketplace where pricing and access adapt per customer type, deployable as a single Docker image.
ReactTypeScriptNode.jsExpressMySQLDocker
Visit agnadeen.shop
allure-id.com
04
Allure Interior Design
Bilingual marketing site + content management system
- Public site + CMS
- Self-service content
- Bilingual EN/AR
Problem
A luxury interior-design studio needed a public portfolio site plus a way for non-technical staff to manage its content.
Architecture
Public React front end paired with an admin CMS dashboard over a Node/Express + MySQL backend, managing services, projects and testimonials. Bilingual EN/AR throughout.
Outcome
A live marketing site the studio updates itself through a custom admin dashboard — no developer in the loop for content.
ReactTypeScriptNode.jsExpressMySQL
Visit allure-id.com

06Technical skills

The toolkit, grouped honestly.

No skill bars or percentages — just what I reach for, organized by where it sits in the stack.

Languages

C#
TypeScript
JavaScript
Python
SQL

Backend & Frameworks

.NET 9
.NET Framework
Node.js
Express
FastAPI
REST APIs

Architecture & Patterns

Clean Architecture
CQRS
Mediator
Dependency Injection
Microservices
OOP
Design Patterns

Frontend

React 18
MUI
Tailwind CSS
Vite
HTML5
CSS3

AI & LLM

AI Agents
Chatbots
RAG Systems
LLM Integration
Groq
Gemini
Cohere
Prompt Engineering
n8n

Data

SQL Server
MySQL
PostgreSQL
pgvector
TypeORM
Redis

Testing & Automation

Selenium
XUnit
Regression Testing

Infra & Exposure

Docker
Git
RabbitMQ
OAuth2 / OIDC
AWS Secrets Manager
Swagger / OpenAPI

07How I build

Clean Architecture, in practice.

The system I work in daily: a layered .NET codebase where dependencies point inward, reads and writes are separate paths, and cross-service work is asynchronous.

Dependencies point inward ↓

React 18 + MUIPresentation — wizards, forms, tables
API / ControllersHTTP boundary — DTOs in, results out
Application (CQRS)Commands, queries & handlers via Mediator
DomainEntities & business rules — no dependencies
InfrastructureEF Core, RabbitMQ, Redis, external services

CreateOrderHandler.cs

// A request/response command — no logic, just intent.
public sealed record CreateOrderCommand(
    Guid CustomerId, OrderDto Payload) : IRequest<Result<Guid>>;

// The handler orchestrates the domain; it depends on interfaces only.
public sealed class CreateOrderHandler
    : IRequestHandler<CreateOrderCommand, Result<Guid>>
{
    private readonly IOrderRepository _repo;
    private readonly IEventBus _bus;

    public CreateOrderHandler(IOrderRepository repo, IEventBus bus)
        => (_repo, _bus) = (repo, bus);

    public async Task<Result<Guid>> Handle(
        CreateOrderCommand cmd, CancellationToken ct)
    {
        var order = Order.Create(cmd.CustomerId, cmd.Payload);

        await _repo.AddAsync(order, ct);
        await _bus.PublishAsync(new OrderCreated(order.Id), ct);

        return Result.Success(order.Id);
    }
}

Dependencies point inward — the domain knows nothing about the framework.
Reads and writes are separate paths (CQRS) handled through a mediator.
Cross-service work is asynchronous over RabbitMQ; Redis absorbs hot reads.
Auth is centralized with OpenIddict (OAuth2 / OIDC); secrets live in AWS Secrets Manager.

08Education & certifications

Engineering foundations.

Education

B.Sc. Communication & Electronics Engineering
2016 – 2021
Delta University for Science and Technology
Graduation project: a real-time driver-drowsiness alert system in Python, using the dlib library for facial recognition to detect driver fatigue and enhance road safety.

Certifications

React Development
Information Technology Institute (ITI)
ASP.NET Core Web API from Scratch (.NET 8)
Udemy
Docker for .NET Developers
Udemy

09Contact

Open to full-time & contract engineering roles.

If you’re hiring for a full-stack or backend role and want someone who ships across the stack, the fastest path is an email or my CV.

Download CV Email me

Software & AI engineer building production systems end to end.

The short version.

Where I've shipped.

Full-Stack Software Engineer

Full-Stack Engineer

National scale, in production.

nafeza.gov.eg

Engineering AI products, end to end.

RAG Service

AI Hub

Products, shipped and live.

Dorosna

21Reve

Agnadeen

Allure Interior Design

The toolkit, grouped honestly.

Languages

Backend & Frameworks

Architecture & Patterns

Frontend

AI & LLM

Data

Testing & Automation

Infra & Exposure

Clean Architecture, in practice.

Engineering foundations.

Education

B.Sc. Communication & Electronics Engineering

Certifications

Open to full-time & contract engineering roles.