Computer Vision Research Engineer, Object Recognition and Pose Estimation - Internship/Fellowship Program
at Magic Leap, Inc.
Computer Vision Reseacher / Computer Vision Research Engineer - Graduate Program
Our company vision is to amplify human potential. Our mission is to deliver enterprise a powerful tool for transformation—an augmented reality platform of great utility and simplicity. Achieving our goals requires passion and dedication. That’s why we’re committed to building and empowering a diverse team of incredibly talented people and fostering an inclusive culture through our values of unity, innovation, and user centricity.
Magic Leap is looking for PhD, Masters and Bachelors students to join our team for a 3 month Summer Program for Underrepresented Students in Engineering, this program is based in Zurich, Switzerland.
You will work fully integrated in Magic Leap’s perception team and help shape the perception systems of our newest generation devices.
Depending on your interest and skills you will work on object recognition and pose estimation in the area of:
Improvement of object pose estimation pipeline in terms of precision and inference speed with state-of-the-art techniques
Investigate object pose estimation pipelines with different input modalities (intensity, depth, IMU)
Temporal object recognition, tracking and pose estimation
Direct pose estimation by leveraging the coarse object reconstruction mesh or CAD
Few-shot learning for object pose estimation with semi-supervised and self-supervised learning
Investigation of customer guided data acquisition to enable better object pose estimation on the limited collected data
Creating mesh reconstruction and automatic pose estimations on new object models by Neural Radiance Field data generation
You have just finished or are finalizing your MSc or PhD degree (preferred) in Computer Science , Robotics, Electrical Engineering or equivalent
Fluent in Python and C/C++
Deep Learning, Computer Vision and 3D geometry knowledge
Projects in Object Recognition, Panoptic Segmentation, Object Pose Estimation in 2D and 3D, and geometric reconstruction
Nice to have
Deep Learning using unsupervised and self-supervised approaches
Geometric understanding and rendering with Neural Readiance Fields (NeRF)