Buildings consume a significant amount of energy, and the exact breakdown depends on a myriad of factors including usage modalities, age, weather, occupants to name a few. The first step to managing their energy usage is understanding where the energy is being consumed, so that it can be managed effecitively. We have worked on making buildings more energy efficient at a campus scale by devising novel sensing methods to detect occupants and their movements and using that information to actually control or actuate building subsystems. We have built and deployed systems to manage IT (e.g. desktops and laptops) energy usage, called Somniloquy and SleepServer. We have also devised wireless, battery powered, occupancy sensors and our own smart energy meter. Using the occupancy information we have actuated the HVAC system of our building as well as managed plug-loads. We are also working on designing and implementing BuildingDepot, a scalable and extensible open-source software system to manage all data related to buildings supported by a library of "connectors" and "apps". There are many exciting avenues of research in the smart-bulding space, including novel sensing systems, at scale networking, data analysis, modeling and actuation. Futhermore, since buildings essentially house human occupants understanding their behavior and motivations, as well as devising the right incentive mechanisms to engage them are important research problems.