About the job
At Intrinsic Robotics, a pioneering initiative from Alphabet, we are on a mission to revolutionize the landscape of industrial robotics. Our belief is rooted in the idea that advancements in artificial intelligence, perception, and simulation will reshape the capabilities of industrial robotics in the near future, with software and data at its foundation.
We aim to make industrial robotics smarter, more accessible, and user-friendly for countless businesses, entrepreneurs, and developers. Our vibrant team, comprised of engineers, roboticists, designers, and technologists, is dedicated to unlocking the creative and economic potential inherent in industrial robotics.
Role
Are you driven by the desire to democratize robotics and make state-of-the-art technology available to all? We are looking for enthusiastic and talented interns to join our innovative team and help propel the future of robotics through this exciting internship opportunity.
Impact of Your Work
- Design and implement a comprehensive dense multi-view reconstruction pipeline using posed images to produce precise 3D ground truth for single-shot reconstruction algorithms.
- Utilize deep learning-based Structure from Motion (SfM) methods to establish a robust benchmark for assessing point-cloud models.
- Evaluate reconstruction quality and, if required, integrate and calibrate additional 3D sensors within the data collection framework to guarantee high fidelity and completeness.
- Work closely with the perception team to validate ground-truth pipelines and facilitate the direct assessment of advanced perception algorithms.
- Enhance internal data tools by setting reliable standards for point-cloud accuracy.
Required Skills for Success
- Pursuing a Bachelor’s degree in Computer Science, Robotics, Computer Vision, or a related discipline (Master's students preferred).
- Strong programming skills in C++ and/or Python for software development and algorithm implementation.
- Familiarity with 3D computer vision concepts, particularly Structure from Motion (SfM), Multi-View Stereo (MVS), or photogrammetry.
- Understanding of point cloud processing, 3D geometry, and linear algebra.
- Experience with 3D vision libraries and tools (e.g., COLMAP, Open3D, PCL, or similar).
- A passion for tackling challenging problems in benchmarking, data evaluation, and perception algorithms.

