lyes arab

I am a data scientist and full‑stack developer with hands‑on experience deploying machine learning solutions to optimize operations and forecasting. I specialize in building end‑to‑end systems that combine predictive models with production APIs and user interfaces.

At Fast Delivery I implemented demand forecasting with Amazon Forecast and SageMaker, integrated TensorFlow models, and deployed an internal AI assistant based on LLaMA 3, while developing NestJS APIs and Next.js/Vue.js frontends.

Previously, I developed predictive and deep learning models for Naftal to improve oil sales forecasts, delivering measurable time and cost savings, and built a CRM web app using Node.js/Pug for Cosider Canalisation to automate resource processes.

I bring strong skills in backend and frontend development, ML model development and deployment, and a pragmatic focus on delivering solutions that reduce costs and improve operational efficiency.