Amirouche MESROUR

Owns a production next-best-action engine end to end, from CP-SAT optimization core to API layer, solving daily under channel, budget, and contact-pressure constraints for ~2M customer portfolios. Shipped a RAG assistant over technical documentation and contributed to TD3-based reinforcement learning agents for renewable energy optimization.

Led development and deployment of a hybrid recommendation engine (LightFM) for a streaming platform with ~150K users and 5K+ titles. Designed a constrained optimization engine for marketing actions across ~2M customer portfolios and built an internal AutoML platform adopted by three production clients.

Built predictive models (propensity, churn, cross-sell, price elasticity) for banking, insurance, and telecom clients. Contributed to internal platforms for audience data monetization and standardized prediction modules.

Taught Statistics 2 and Simulation, and mentored students in statistical methods and probabilistic modeling.

Modeled constrained occupancy of port terminal chains to optimize container ship loading and unloading flows using operations research and simulation.

M.Sc. in Mathematics with a concentration in Statistics from Université du Québec à Montréal.

M.Sc. in Applied Mathematics from the University of Béjaïa, Algeria.