Development of an autonomous defense system: prediction of DDoS attacks and automatic recommendation of countermeasures (RL & IA decision-making)

CONTEXT Previous work has enabled the development of DDoS attack prediction models with over 99% accuracy. However, prediction alone is not enough: it is also necessary to quickly decide on the most suitable countermeasures to neutralize detected attacks. The goal of this internship is to design an intelligent automatic mitigation module combining decision-making AI and SOC knowledge. INTERNSHIP OBJECTIVES Design an autonomous defense system that recommends and simulates the most suitable DDoS countermeasures in real-time, based on the predictions of the detection model. TASKS REQUIRED • Model DDoS mitigation as a sequential decision process (MDP) • Develop an intelligent decision engine (Reinforcement Learning, contextual bandits, hybrid expert systems) • Build a network simulator to evaluate defense strategies • Evaluate defense performance (mitigation rate, latency, false positives) • Integrate the module in real-time with the existing detection pipeline DELIVERABLES • Operational decision model • Complete real-time pipeline + REST API • SOC dashboard integrating recommendation visualization • Scientific report