IYAD KHALIL

Developed an AI-powered SaaS e-commerce platform supporting multi-vendor architecture with automated product handling and real-time order routing. Built a hybrid recommendation engine combining collaborative filtering and content similarity using WebGPU for fast client-side inference. Integrated OpenAI GPT-4 for multilingual product description generation and dynamic translation with SEO-optimized prompt engineering. Applied Hugging Face SegFormer for real-time background removal and image enhancement. Architected a modular stack using React, Supabase Edge Functions, PostgreSQL with RLS, and Stripe for secure payments and vendor-level isolation.

Developed a hybrid LSTM + XGBoost model for multivariate forecasting of pharmacy revenue based on enriched time series data. Engineered 25+ explanatory features including sales history, holidays, seasonality, patient behavior, volatility, long-term trends, and more. Achieved a prediction accuracy of 93.21% with SMAPE less than 10% on real-world test sets. Designed a secure RESTful API using FastAPI, containerized with Docker, served with NGINX, and deployed on a Linux Azure VM. Integrated an intelligent conversational agent using Azure OpenAI and LangChain for personalized recommendations. Implemented a robust MLOps architecture with model versioning, virtual environments (Conda), API monitoring (Postman), and reproducible pipelines.

Automated end-to-end testing for a web application using Cypress. Covered most platform features with reusable and maintainable test scripts. Collaborated with developers to detect bugs and improve overall product quality.