Water and wastewater treatment are energy intensive processes that account for 3% of the energy load in the United States. As water resource recovery facilities strive to become more sustainable, the need to implement energy efficient treatment technologies is further motivated by global challenges such as climate change, population growth, and aging infrastructure. This dissertation examines how an energy-efficient wastewater treatment technology, the membrane aerated biofilm reactor (MABR), can be designed and operated to support the growth and retention of anaerobic ammonia oxidizing (anammox) bacteria to efficiently remove nitrogen in mainstream wastewater. This research addresses practical questions focused on optimizing reactor startup, identifying the reactor configuration which maximizes nitrogen removal, and determining the overall performance of a reactors-in-series configuration. A unique MABR startup method that was designed to limit the growth of nitrite oxidizing bacteria (NOB) and encourage the growth of anammox bacteria (AMX) was evaluated against a traditional startup. For this study, an MABR was started anaerobically with nitrite added to the influent to enrich AMX. Once AMX was established, air was added through the membrane and exogenous nitrite addition was stopped. Our hypothesis was that having an established AMX population in the biofilm when air was added would suppress the growth of NOB. Computer modeling was used to compare the novel startup to a traditional startup that added air immediately to the MABR. The results show that NOB suppression and total nitrogen removal performance were similar for both strategies, which suggests that additional NOB suppression strategies may be needed for mainstream nitrogen removal. Simulation experiments were used to evaluate the impact of reactor configuration on overall nitrogen removal performance. MABR research that incorporates anammox has thus far primarily focused on one-stage (MABR only) configurations rather than hybrid (MABR + suspended growth) or two-stage (MABR + traditional biofilm reactor) configurations.1–7 A simulation study was conducted to evaluate and compare the performance of these three configurations. The results show that the two-stage configuration removed twenty percent more nitrogen than the one-stage and hybrid configurations. The two-stage configuration also had increased metabolic flexibility in that it supported both anammox and heterotrophic denitrification. Furthermore, the impact of placing MABR units in series (staging) on nitrogen removal performance, NOB suppression, and operation was investigated. MABR performance was simulated at different influent total ammonia concentrations and loadings. Using this performance data, a design methodology was created to determine the membrane surface area needed to achieve a desired performance in staged MABRs. The staging methodology’s validity was confirmed against treatment trains with different distributions of membrane surface area in three staged reactors. Finally, the performance of an experimental MABR that underwent periods of anammox enrichment, intermittent aeration, and continuous aeration was evaluated. Reactor results show that the enriched AMX were not able to outcompete NOB, and the performance significantly decreased when air was added to the reactor. The reactor had a low nitrogen removal rate throughout both aeration periods and was not significantly affected by switching from intermittent to continous aeration. The results of this dissertation show that AMX can be incorporated into MABRs for mainstream nitrogen removal. However, additional research on NOB suppression is required. Ultimately, this research can support the practical scale-up of anammox-driven processes in MABRs and will hopefully increase adoption of the MABR technology for nitrogen removal during wastewater treatment.