Reinforcement Learning Intern

A practical internship focused on extending an existing soil-flow optimization tool that combines discrete-event simulation with reinforcement learning for dike construction projects. The intern will adapt the model to a new project, design reward functions for KPIs such as cost and emissions, and train and evaluate Maskable PPO agents on real project data. The work is done alongside engineers and data scientists, with strong emphasis on experiment tracking, validation and reproducible documentation.