Quantum dots are a promising platform to realize practical quantum computing. However, before they can be used as qubits, quantum dots must be carefully tuned to the correct regime in the voltage space to trap individual electrons. Moreover, realizable quantum computing requires tuning of large arrays, which translates to a significant increase in the number of parameters that need to be controlled and calibrated. This necessitates the development of robust and automated methods to bring the device into an operational state. Building upon previous ray-based tuning methods, we explore the utility of the ray-based measurements and maximum entropy to accurately model the quantum dot system as a multivariate gaussian, leading to a statistical manifold with Riemannian metric structure. The tools of information geometry are used to derive controls along energy\entropy gradients to provide a coordinate basis for targeted control of chemical potential and tunnel coupling of the quantum dots. Our work is an important step in establishing reliable fine-tuning methods for calibrating quantum dots to work as qubits.