Tech Stack

Description

As an intern, you will develop a dense multi-view reconstruction pipeline using posed images to generate accurate 3D ground truth. You'll apply deep learning-based Structure from Motion techniques, collaborate with the perception team, and contribute to internal data tools for point-cloud accuracy.

Requirements

• Bachelor's student in CS, Robotics, CV or related field (Master's preferred)
• Proficiency in C++ and/or Python
• Experience with 3D computer vision: SfM, MVS, or photogrammetry
• Familiarity with point cloud processing, 3D geometry, linear algebra
• Experience with 3D vision libraries (e.g., COLMAP, Open3D, PCL)

Responsibilities

• Develop and implement dense multi-view reconstruction pipeline
• Apply deep learning based SfM techniques for benchmark creation
• Analyze reconstruction quality and integrate additional 3D sensors
• Collaborate with perception team to validate ground-truth pipelines
• Contribute to internal data tools for point-cloud accuracy standards