Explore how we turn complex ideas into scalable digital solutions. Each project is a partnership built on strategy, design, and precision engineering.
9 projects
A digital platform that helps Algerians import vehicles and spare parts from Europe, China, and Dubai — with full confidence and no financial intermediary. Cabacar connects buyers with certified garages abroad, organizes import documents, translates conversations in real time (Arabic, French, English), and tracks each order step by step until registration in Algeria.
A medical learning platform with video courses and QCM, now with AI chat for PDFs and images (AI-generates mind maps, MCQs, and flashcards), integrated 3D anatomy models, and low-latency video streaming plus a full admin panel for users, courses, quizzes, and content.
An automated order-fulfillment platform built for e-commerce scale. The system synchronizes data between business owners and confirmation agents, routing validated orders to shipping partners via a centralized management interface.
A fitness management web application featuring automated PDF generation for training and nutrition. Coaches track client progress and deliver structured, real-food meal plans tailored to schedules and tastes.
Tool that helps Shopify stores and agencies monitor Google Merchant Center problems, fix product/feed issues faster, reduce disapprovals, track suspension risks, and prepare evidence/reporting.
BodyVibe is a comprehensive, all-in-one health and fitness platform. It bridges the gap between wellness and technology by allowing users to book paid consultations with verified doctors and fitness coaches, with a marketplace, real-time chat, and an AI assistant for safe gym machine usage.
An immersive 3D online e-commerce shop that feels like a real-life marketplace.
A real-time collaborative whiteboard with audio and text messaging to enhance online learning experiences.
A production-grade image classification system engineered from first principles — no TensorFlow, no PyTorch, no recurring API fees. We built and benchmarked three custom-implemented architectures (Linear regression with Ridge regularization, Multi-Layer Perceptron with full backpropagation, and Radial Basis Function with Lloyd's K-Means clustering), then shipped the winner as a memory-safe Rust binary exposed to Python through ctypes and served over Flask. The result: real-time inference in a few-megabyte footprint that runs on edge devices, internal servers, or embedded hardware — fully owned, fully tunable, with zero vendor lock-in.
Results vary by project. Metrics reflect outcomes achieved through our collaboration.
