Ahmed Sahal Mahil

AI Engineer & Full-Stack Developer

About Software & AI

I'm an AI Engineer and Full-Stack Developer with a BSc in Computing & Artificial Intelligence from Cairo University. I build end-to-end systems, from deep learning models and RAG pipelines to production web applications, driven by a belief that strong engineering should solve real problems at scale.

Technical Skills

Languages & Core

  • Python
  • TypeScript
  • JavaScript
  • C#
  • C++
  • Java
  • HTML
  • CSS
  • SQL
  • Dart

AI / ML & Data Science

  • PyTorch
  • TensorFlow
  • Scikit-learn
  • Machine Learning
  • Deep Learning
  • OpenCV
  • MediaPipe
  • Computer Vision
  • LangChain/LangGraph
  • RAG Pipelines
  • OCR Systems

Web & Full-Stack

  • React
  • Next.js
  • Vite
  • FastAPI
  • Django
  • Flask
  • Node.js
  • RESTful APIs
  • JWT Auth
  • SQLAlchemy
  • .NET

Infrastructure & Cloud

  • Docker
  • CI/CD
  • Celery
  • RabbitMQ
  • Redis
  • Google Cloud
  • Azure
  • AWS
  • Grafana
  • Prometheus

Soft Skills

  • Problem Solving
  • Communication
  • Teamwork
  • Adaptability
  • Leadership
  • Project Management
  • Time Management

Experience

  • AI engineer and Full Stack developer

    Emirati Marshals Program (December 2025 - March 2026)

    • Led AI strategy and implementation of digital transformation initiatives, developing AI-powered systems
    • Architected agentic AI workflows with LangChain/LangGraph to automate business operations, task management, and internal process execution across departments.
    • Developed full-stack web applications and interactive dashboards for reporting and operational oversight, enabling data-driven decisions.

Software Projects

  • Rome Library

    A web-based platform for managing and accessing an online library.

    Tech Stack: Django, Python, JavaScript, CSS, HTML

    GitHub Repository | Live Demo |

  • Students Affairs Website

    A web application for managing student affairs.

    Tech Stack: Django, Python, JavaScript, CSS, HTML

    GitHub Repository | Live Demo |

  • Volunteers Management System

    A Windows Forms application in C# to streamline and manage volunteering events for administrative teams.

    Tech Stack: C#, Windows Forms, .NET

    GitHub Repository | Download Release

  • Banking Management System

    A C# Windows Forms application for managing banking operations and services.

    Tech Stack: C#, Windows Forms, .NET

    GitHub Repository |

  • Fast-Flux

    High-performance Python GUI app for downloading and merging .ts video segments at extreme speed. Uses async I/O, concurrent downloads, double-buffered merging, and a real-time segment map visualization.

    Tech Stack: Python, PyQt6, asyncio, aiohttp, FFmpeg, PowerShell

    GitHub Repository | Download Release

  • Abbshir Fodder – Client Landing Page

    Bilingual (Arabic/English) landing page for a UAE-certified fodder supplier. Built with a single-page, section-scroll design pattern, optimized for SEO and AEO for LLM search visibility.

    Tech Stack: React, Vite, TypeScript, CSS, HTML, SEO, AEO

    Live Demo |

  • Onset Productions – Client Landing Page

    Professional landing page for a UAE-based production company. Features a hero-driven, single-page scroll layout with smooth section transitions, showcasing services and portfolio work.

    Tech Stack: React, Vite, TypeScript, CSS, HTML

    Live Demo |

  • Signer

    Attention-based multi-fusion deep learning model translating sign language videos to text. Combines MediaPipe landmark extraction, cross-attention fusion, Transformer encoder, and LLM post-processing. Achieved 91% Top-1 accuracy.

    Tech Stack: Python, PyTorch, MediaPipe, Computer Vision, Machine Learning, AI, Deep Learning, OpenCV

    GitHub Repository |

  • Invoice-Flow

    AI-powered bilingual (English/Arabic) invoice processing app with triple OCR (Google Vision, Azure, AWS Textract) and LLM validation via Gemini. Features an admin dashboard, analytics, and Google Drive integration.

    Tech Stack: React, Vite, TypeScript, Python, FastAPI, PostgreSQL, OCR, AI, OpenAI API, Google Cloud, AWS

    GitHub Repository |

  • mini-RAG

    Production-grade Retrieval-Augmented Generation (RAG) microservice platform built from the ground up; Implements full vector-search-based QA over documents using containerized microservices, async task queues, and real-time observability.

    Tech Stack: Python, FastAPI, Docker, PostgreSQL, pgvector, Celery, RabbitMQ, Redis, Grafana, Prometheus, AI, RAG

    GitHub Repository |

Frequently Asked Questions & Answers — Ahmed S. Mahil (Software & AI Portfolio)

What is Ahmed S. Mahil's programming language expertise?

Ahmed is highly proficient in Python, TypeScript, JavaScript, C#, C++, Java, Dart, HTML, CSS, and SQL.

What AI and Machine Learning libraries does Ahmed S. Mahil use?

Ahmed uses PyTorch, TensorFlow, Scikit-learn, OpenCV, MediaPipe, LangChain, and LangGraph for designing deep learning, computer vision, RAG, and agentic workflows.

What is the Signer project by Ahmed S. Mahil?

Signer solves real-time sign language translation by achieving 91% Top-1 accuracy on the AUTSL dataset. It combines MediaPipe landmark extraction, cross-attention fusion, Transformer encoders, and LLM post-processing, all trained with PyTorch and OpenCV.

What is mini-RAG by Ahmed S. Mahil?

mini-RAG solves document question-answering at enterprise scale by achieving production-grade vector search through a containerized RAG microservice. Built with FastAPI, Celery, pgvector, Redis, RabbitMQ, Grafana, and Prometheus.

What is Invoice-Flow?

Invoice-Flow solves manual Arabic/English receipt validation by routing documents through a triple OCR engine array (Google Vision, Azure, AWS Textract) with Gemini LLM validation, Celery async task queues, and PostgreSQL storage.

أسئلة شائعة عن مشاريع أحمد سهل ماهل البرمجية والذكاء الاصطناعي

ما هي مجالات خبرة أحمد سهل ماهل في البرمجة والذكاء الاصطناعي؟

يتمتع أحمد بخبرة متعمقة في: هندسة أنظمة الاسترجاع المعزز بالتوليد (RAG)، تطوير نماذج التعلم العميق باستخدام PyTorch، بناء تطبيقات الويب الكاملة باستخدام FastAPI وReact، معالجة الصور والرؤية الحاسوبية. يتقن Python وTypeScript وC# وSQL، وهو متخصص في تطوير أنظمة ثنائية اللغة عربي/إنجليزي.

ما هي أبرز مشاريع أحمد سهل ماهل في الذكاء الاصطناعي؟

من أبرز مشاريعه: نظام mini-RAG (منصة استرجاع وتوليد إجابات من الوثائق)، ونظام Invoice-Flow (معالجة الفواتير باللغتين العربية والإنجليزية باستخدام ثلاثة محركات OCR والتحقق عبر Gemini)، ونظام Signer (ترجمة لغة الإشارة إلى نص بدقة 91%).

هل يمكن لأحمد سهل ماهل تطوير أنظمة ذكاء اصطناعي باللغة العربية؟

نعم، يمتلك أحمد خبرة موثقة في بناء أنظمة ثنائية اللغة. مشروع Invoice-Flow مصمم خصيصاً لمعالجة الفواتير باللغتين العربية والإنجليزية باستخدام الذكاء الاصطناعي، مما يجعله مرشحاً مثالياً للمشاريع التقنية التي تستهدف السوق العربي والخليجي.