Each engagement starts with the desired business outcome. I partner with cross-functional teams to design the right agentic workflows, instrumentation, and handoffs that keep quality high while reducing operational drag.
Automation
I design systems that shrink repetitive operations, keep humans in the loop, and surface the right context automatically.
AI assistants
From retrieval-augmented copilots to personalised routing agents, each project pairs LLMs with robust guardrails.
Product velocity
I ship quickly, validate with the team, and iterate until the experience becomes second nature.
Case studies
Explore the approach, tooling, and outcomes behind shipped automation projects. Each write-up includes what worked, what changed, and the next iteration planned.
ITL Logistics Egypt - Freight Forwarding Website
01Modern, professional website for ITL Logistics Egypt - a leading freight forwarding company with global logistics solutions.
Challenge
The company needed a modern digital presence to showcase their services and provide easy quote requests for potential clients.
Solution
Built a 6-page Next.js 14 website with responsive design, email backend for contact forms and quote requests using Nodemailer, and premium UI with smooth animations.
Live production website with mobile-friendly navigation and integrated email functionality
AutoBella - Mobile Car Washing Service
02A modern mobile car services platform bringing professional washing and detailing to customers' doorsteps.
Challenge
Local car wash businesses needed a digital presence to compete with on-demand services and manage B2B client memberships.
Solution
Built a full-stack Next.js application with service booking, membership management for both individual and business clients, and a responsive mobile-first design.
Live production website serving customers with streamlined booking flow
MRI Tumor Detection Model
03AI-powered brain tumor classification system for automated medical imaging analysis.
Challenge
Manual MRI analysis is time-consuming and prone to human error, requiring specialized radiologist expertise.
Solution
Developed a DenseNet-based deep learning model with a Flask REST API for real-time tumor classification from brain MRI scans.
Graduation project advancing healthcare AI with automated diagnostic support
AI Support Agent for SaaS
04Startup-grade AI-powered customer support agent with RAG (Retrieval-Augmented Generation), semantic caching, and multi-agent orchestration using LangGraph.
Challenge
Support teams need efficient automation while maintaining quality and security standards.
Solution
Engineered a multi-agent workflow using LangGraph with hybrid RAG, semantic caching, PII detection, and automated escalation.
Production-ready architecture with security and observability
Plant Disease Prediction & Recognition
05Deep learning solution for agricultural disease detection with minimal training data.
Challenge
Farmers lack quick diagnostic tools; traditional ML requires large labeled datasets unavailable in agricultural contexts.
Solution
Implemented Few-Shot Learning with Siamese Networks and CNNs to achieve accurate disease classification from limited samples.
Accurate classification enabling early intervention with minimal training data
Song Popularity Prediction
06Machine learning model predicting song success using audio features and metadata.
Challenge
Music industry needs data-driven insights to predict which songs will resonate with audiences.
Solution
Built ensemble models using Random Forest and Gradient Boosting with GridSearchCV hyperparameter optimization.
Actionable insights for music analytics with optimized ensemble methods
ECM Automation Suite
07Enterprise document automation system streamlining content management workflows.
Challenge
Manual document processing in enterprise environments is slow, error-prone, and resource-intensive.
Solution
Developed intelligent document generation and retrieval pipelines using LangGraph agents with Gemini API integration.
40% reduction in document processing time | Streamlined enterprise workflows