Self-evolving engineering curricula: A Reinforcement Learning architecture to align academic training with industry demands in developing countries

Abstract

Engineering education in developing countries faces a widening gap between academic curricula and labor market needs, as traditional revision processes are too slow to keep pace with rapid technological change. This paper proposes a conceptual, self-evolving curricular architecture based on RL, formalizing academic programs as a Markov Decision Process in which curricular adjustments are actions, academic and employability indicators are state variables, and a multi-criteria reward function balances performance, employability, industrial alignment, and budget constraints. An Explainable AI layer and a human-governed workflow support institutional trust. As a proof-of-concept contribution, no deployment, dataset, or simulation has been carried out; the paper instead formalizes the problem, details the architecture, and outlines a validation protocol for future empirical work.