Paulo Granja

At Jumia, Africa's largest e-commerce platform, I lead the AI/ML roadmap for millions of daily users across 11 countries. I architected and deployed an end-to-end AWS AI/ML platform supporting recommendation, dynamic pricing, search ranking, and demand forecasting.

I built a production recommendation engine that increased cross-sell revenue and average order value, along with pricing and search-ranking systems that improved conversion, competitiveness, relevance, and CTR. I also delivered demand forecasting models that reduced overstock and stockout costs across multiple markets.

I bring Generative AI into production through RAG-based support assistants and LLM-powered product-description generation, reducing operational costs. My MLOps work includes feature stores, model registries, automated retraining, A/B testing, Terraform, CI/CD, SageMaker, MLflow, Docker, and Airflow.

Previously, I led data science teams and optimization initiatives at ITIM and delivered predictive analytics, marketing models, forecasting, and Hadoop-based reporting at Sonae MC. I enjoy turning complex business problems into scalable AI solutions with measurable P&L impact while mentoring teams and aligning technical roadmaps with product and engineering leadership.