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.
AI Engineer & Full-Stack Developer
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.
Emirati Marshals Program (December 2025 - March 2026)
A web-based platform for managing and accessing an online library.
Tech Stack: Django, Python, JavaScript, CSS, HTML
A web application for managing student affairs.
Tech Stack: Django, Python, JavaScript, CSS, HTML
A Windows Forms application in C# to streamline and manage volunteering events for administrative teams.
Tech Stack: C#, Windows Forms, .NET
A C# Windows Forms application for managing banking operations and services.
Tech Stack: C#, Windows Forms, .NET
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
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
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
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
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
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
Ahmed is highly proficient in Python, TypeScript, JavaScript, C#, C++, Java, Dart, HTML, CSS, and SQL.
Ahmed uses PyTorch, TensorFlow, Scikit-learn, OpenCV, MediaPipe, LangChain, and LangGraph for designing deep learning, computer vision, RAG, and agentic workflows.
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.
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.
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.
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.
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%).
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.