Leading SLAM, Persistence and Online Calibration teams for Magic Leap's current and future devices.

Leading a cross site international team to deliver state-of-the-art perception algorithms for Magic Leap's upcoming mixed reality glasses. Drive research efforts for new algorithms and solutions for enterprise customers. Coordinating with various software and hardware teams over the whole product development cycle from negotiating product requirements to incorporating customer feedback.Responsible individual for 6DoF tracking of the user's head in 3D space.

At Dacuda I worked as technical lead of 3D reconstruction and tracking on several projects in the field of mobile vision together with major technology companies. Dacuda's goal was to enable inside-out tracking and 3D reconstruction on standard smartphones.Most recently, I was the team lead for Dacuda's SLAM based face scanning technology enabling 3D printouts of personal figurines and the creation of virtual avatars, presented at Siggraph'16 and CES'17. Previously, our team was working on increasing the robustness, the accuracy and the latency of Dacuda's tracking algorithm to bring 6Dof tracking for mobile VR to the market.Moreover, in the beginning of 2015 I led the development of a panorama application supporting arbitrary camera movements, which is licensed to major smartphone manufactures and integrated into their platforms.

SMI develops and produces eye-tracking devices and it is one of the technical leaders in the field. First and foremost I worked on the integration of SMI's eye-tracking technology into head mounted displays (Oculus DK2). In this project I was responsible for developing a new methodology for calibrating the optical system including camera and lenses. Moreover, I extended SMI's eye-tracking algorithms to cope with the lens distortions of the HMD. In another project I ported the software for SMI's remote eye-tracking devices to Android.

As the leader of the research group “Industrial and Medical Imaging” I carried the responsibility for doctoral students and federal German as well as EU research projects (SFB 871, EU CleanSky), which allowed me to improve my leadership qualities. Additionally, I taught the course Industrial Imaging and developed new research projects including the application for external funding.Two projects with our industrial partners in the area of scene analysis were most interesting to me. The first one aimed at developing smart operating room equipment with the help of 3D sensors and the second was concerned with the inspection of roads also based on low-cost 3D sensors.

In the research training group 1564 “Imaging New Modalities” funded by the German Research Foundation, I conducted research mostly on 2D/3D imaging. Together with my colleagues I developed a monocular 2D/3D camera based on Time-of-Flight imaging chips. In particular, I was involved in the following major projects:- 2013: Joint project with Elektroniksystem- und Logistik-GmbH (ESG). Development of a prototype of a landing assistance system for helicopters based on 2D/3D imaging. Hardware design, driver development, testing as well as the documentation were my responsibilities in this project.- 2012-2013: LaotSe - Airport Runway Inspection with Multi-Modal Interconnected Sensors. In this research project a safety system based on RADAR and 2D/3D imaging was developed in cooperation with the Fraunhofer institutes FHR and FKIE. I designed, implemented and evaluated a long-range 2D/3D imaging system capable of measuring in up to 150 meters distance.- 2011-2012: MOVEDETECT - Secure Detection, Localization and Classification in Wireless Sensor Networks. Together with two other universities and two companies, a distributed sensor system for the protection of critical infrastructure was designed and implemented funded by the German federal police. I was partly responsible for the software development and the testing of the sensor nodes (micro-controller programming). Moreover, I developed a system to map the sensor nodes using differential GPS and designed algorithms for the detection, classification and tracking of intruding objects operating on an embedded Linux platform.