The genome sequencing programs have identified hundreds of thousands of genetic and proteomic targets for which there are presently no ascribed functions. The challenge for researchers now is to characterize them, as well as identify and characterize their natural variants. Historically, this has meant studying each individual target separately. However, due to the recent development of multi-analyte microarray devices, these characterizations can be performed in a combinatorial manner in which a single experiment provides information on thousands of targets at a time. In the past decade, microarray technology has settled in on two major designs. The first entails spotting individual receptor types onto a functionalized glass substrate. This is a simple and inexpensive process; however, due to the limited resolution of the mechanical devices used to do the spotting, the densities of these arrays are relatively low. Moreover, receptor preparation requires substantial time and effort. The second variety of microarray uses photolithographic techniques adapted from the semi-conductor industry to chemically synthesize the receptor elements in situ on the sensing surface. Because lithographic patterning is spatially very precise, these arrays achieve very high densities, with as many as one million features per square centimeter. Although these arrays obviate the necessity for laborious "off chip" probe preparation, they are expensive to produce and are limited to two types of receptors (oligonucleotides and peptides). This dissertation presents the development work performed on a hydrogel-based biosensor platform which provides a high density and low cost alternative to the two aforementioned designs. The array features are fabricated lithographically from a liquid pre-polymer doped with biologically active sensing elements at sizes as small as 50[micrometer]. Each of the feature types is uniquely shaped, which enables the features to be mass-produced in batches, pooled together and then assembled into randomly ordered arrays using highly-parallelized self-assembly techniques. The three-dimensional hydrogel features accommodate a wide variety of sensing elements, such as enzymes, antibodies and cells, which cannot be deployed using the traditional designs. This dissertation presents methods developed to integrate cellular and oligonucleotide sensing elements into the hydrogel features which preserve their biological activity and optimize the sensor's performance. / text