This document summarizes research on the architecture and performance of Aardvark, a large-scale social search engine. Aardvark crawls social media profiles and indexes the interests and expertise of its users to answer questions by connecting askers to relevant answerers in its network. It ranks potential responders based on their topic expertise, social connections, and availability. Evaluation shows Aardvark can find answers to a high percentage of questions in under 3 words and 3.5 responses on average, outperforming Google for certain types of social questions.