Quantum spin-1/2 networks are highly relevant for quantum technologies with applications ranging from quantum spintronics for computing, simulation and networking to magnetic resonance imaging and spectroscopy in healthcare. This project investigates the use of reinforcement learning algorithms to control the dynamics of such networks to implement information transfer using energy landscape shaping. Results are compared with L-BFGS controllers. There is no doubt that under ideal circumstances L-BFGS will perform the best amongst the tested in this paper. However, as soon as input Hamiltonian is distorted (and in real experiments there will be inevitable noise inthe system), the controls are no longer robust and therefore target fidelity can be no longer achieved. Reinforcement learning methods allow the controllers to be robust under noise simulation as well.