About | Manohar Kuse

Scholar GitHub LinkedIn YouTube

I am currently working at P8S, a Switzerland-based technology company. At P8S, I focus on shipping practical perception and robotics software with production-grade reliability.

In the past, I worked as a Senior Software Engineer at Magic Leap on SLAM-related headpose topics in AR devices. At Magic Leap, I learned how to balance algorithm quality with latency, robustness, and long-term maintainability in real products.

In the past, I was employed as a Robotics Navigation Engineer at Rovenso SA (Apr 2020-Nov 2021). I worked on ground robot controllers for docking motion and other robotic navigation topics. Rovenso is a startup incubated at EPFL and based near Fribourg, Switzerland.

I completed my PhD in Jan 2020 from the UAV Group of the Robotics Institute (RI) at HKUST, Hong Kong. My supervisor was Prof. Shaojie Shen. My thesis work included a CNN-based place recognition system for online recovery from kidnap in visual-inertial SLAM. A project overview is available on GitHub.

I have hands-on experience with 3D geometry, stereo vision, place recognition, and object detection. I mostly program in C++ and Python, and regularly use OpenCV, ROS, Eigen, Ceres Solver, GTSAM, and Keras/TensorFlow.

Before joining the PhD program, I spent one year doing research internships at CERN, University of Warwick (UK), and SenoCAD Research (Netherlands). I completed my B.Tech at LNMIIT in Communication and Computer Engineering in 2012.