Antimicrobial resistance is a major public health concern that is embodied by the diminishing utility of available antibiotics. Without antibiotics, infections can be deadly and thus many modern medical procedures will be deemed too risky; reminiscent of an era when such procedures did not exist. Antimicrobial resistance is accelerated by improper use of antibiotics and requires action in the form of global implementation of antibiotic stewardship practices. In clinical settings, where antibiotic resistant infections can develop and spread relatively easily, the misuse of antibiotics is driven by a lack of rapid access to patient specific knowledge such as proof of if the infection actually exists, and if so which antibiotic therapies are appropriate. Determining the appropriate antibiotic for a given infection, through antimicrobial susceptibility testing (AST), currently takes a number of days due to the lengthy culturing steps to obtain large amounts of isolated bacteria and the reliance on insensitive detection systems for cultures of hundreds of microliters in scale. Research efforts have utilized microfluidics to perform AST in only a number of hours using more sensitive readouts and on the nanoliter scale, however have been either too complicated for clinical practicality, lack proper multiplexing, or heavily dependent on external equipment, making them costly and/or lowering their throughput. We propose a method which implements dried antibiotic gradients and a fluorescent metabolic marker within arrays of nanoliter wells for rapid, simple, and accurate AST on a chip. Nanoliter testing volumes reduces the need for lengthy culturing by enabling the use of fewer bacteria while the fluorescent metabolic marker provides increased assay sensitivity and faster AST results. Sample dispersion into hundreds of identical nanoliter wells provides for a more robust assay by quarantining contaminations into individual wells. High surface area to volume microchannels promote ample oxygen flux and rapid bacterial growth and AST. In this thesis, we demonstrate the efficacy of our assay on a number of clinical isolates and clinical urine samples. We develop a lyophilization protocol for pre-loading cartridges with antibiotics for improved usability. We develop a novel multiplexing solution by discretizing a chemical gradient across our arrays of nanoliter wells. Lastly, we investigate the potential of a unique immunomagnetic capture system to bypass the blood culture step for blood samples with suspected fungemia. This method, in combination with a rapid AST system can save 2-6 days off the traditional diagnostic timeline.