While current research focus is model-based generation of optimal robotic locomotion, my interests are much broader. Starting of from providing insight into how to use system dynamics for efficiently moving a robot around, I also worked on feed-back control on high- and low-level of the technologically most advanced humanoid robot. Along the way I created the tools required and shared them among colleges such as simulation software and real-time capable implementation of dynamics computations. Continuous integration of software and hardware components is a reality in robotics, which I experienced when combining an existing state estimation based on stereo vision and inertial sensing with existing feed-back control code. Communication across engineering domains is very important to find a system-level solution, which is my personal goal. Particularly relevant is this in my work of build a testbed for evaluating new series elastic drive concepts for robust and high performance locomotion. Another cross-domain aspect of new robotic applications is perception of the environment which I successfully integrated in multiple projects.