Over the last two decades, astronomers have developed the ability to detect exoplanets, planets orbiting other stars in the Milky Way. Today, thousands of these exoplanets are known, many detected through the subtle acceleration of a host star imparted by its orbiting planet(s). As the measurement techniques that are used to detect exoplanets through these stellar velocity variations have improved, progressively less massive planets have been detected. Over the next decade, a key goal in the field is to develop instruments capable of detecting an Earth-mass planet orbiting a Sun-like star with a period of one year. While detecting such an Earth twin is beyond our current technological capabilities, new technologies may make this soon possible. This dissertation explores several ways in which we can improve upon spectrograph designs to achieve the goal of detecting Earth twins by combining commercially available technologies to significantly reduce the cost of building high-precision instruments designed to detect exoplanets. I will describe the installation, setup, and automation of the MINERVA observatory at the Fred Lawrence Whipple Observatory located on Mt. Hopkins, Arizona. This facility includes MINERVA-Red, a telescope and instrument dedicated to the discovery of planets around the nearest low mass stars. I will describe the motivation, design, construction, and limitations of the current MINERVA-Red spectrograph system, which is designed around single-mode fibers and high-speed guiding, along with its low-cost environmental control system that provides milliKelvin temperature stability for the instrument. I will discuss the current limitations of the MINERVA-Red system and how in the future, single-mode fiber-fed spectrographs may play a key role in the ongoing quest to detect Earth-like planets.