High-performance semiconductor optoelectronics such as perovskites have\nhigh-dimensional and vast composition spaces that govern the performance\nproperties of the material. To cost-effectively search these composition\nspaces, we utilize a high-throughput experimentation method of rapidly printing\ndiscrete droplets via inkjet deposition, in which each droplet is comprised of\na unique permutation of semiconductor materials. However, inkjet printer\nsystems are not optimized to run high-throughput experimentation on\nsemiconductor materials. Thus, in this work, we develop a computer\nvision-driven Bayesian optimization framework for optimizing the deposited\ndroplet structures from an inkjet printer such that it is tuned to perform\nhigh-throughput experimentation on semiconductor materials. The goal of this\nframework is to tune to the hardware conditions of the inkjet printer in the\nshortest amount of time using the fewest number of droplet samples such that we\nminimize the time and resources spent on setting the system up for material\ndiscovery applications. We demonstrate convergence on optimum inkjet hardware\nconditions in 10 minutes using Bayesian optimization of computer vision-scored\ndroplet structures. We compare our Bayesian optimization results with\nstochastic gradient descent.\n