Indoor 3D Vision Researcher

Indoor 3D Vision Researcher

At Geomagical Labs

Date Posted: 11 June, 2021


Geomagical Labs is a visual AI research & development lab in the San Francisco Bay Area, applying novel computer vision technologies to an exciting consumer application to be launched to hundreds of millions of users by a beloved global brand.

We have an opening for a 3D computer vision research scientist with strong academics and substantial experience in 3D reconstruction and multi-view geometry, to join our core startup team of computer vision specialists. In this role, you will work with our team of vision scientists in an applied R&D role, to innovate novel indoor 3D reconstruction technologies for our mobile and cloud-based applications, fusing multiview geometry (SFM/MVS/SLAM), and deep scene perception neural networks with unique priors and side-channel constraints. This could be a dream job for highly-motivated applied researchers who want to see their work used around the globe.

Candidates must have deep proficiency in 3D reconstruction methods including substantial expertise with sparse and dense perspective 3D vision including multiple view geometry, structure from motion, MVS, SLAM. This position requires strong theoretical foundations, excellent software development skills, and sound scientific method. Candidates must also be adventurous and enthusiastically excited about an entrepreneurial innovation experience.

Our headquarters is located in the heart of Silicon Valley in downtown Mountain View, California. We plan to open a satellite office in central Seattle in 2021. Working from one of our office locations is preferred (once public health conditions permits), but experienced candidates who are able to work effectively while remote, will be considered for fully remote work.


  • Strong academic foundations in 3D computer vision (recent Ph.D. from competitive university/lab preferred).
  • Strong grasp of the theory & practice of 3D multiview geometry & photogrammetry.
  • Experience engineering multiview pose estimation systems (SFM, SLAM).
  • Experience developing algorithms for dense 3D reconstruction (MVS).
  • Experience using diverse 3D scene representations (point clouds, voxels, TSDF, etc.).
  • Experience with RGBD fusion algorithms.
  • Familiar with state-of-the-art neural network methods for multi-view geometry.
  • Excellent C++ software development skills in Linux environments.
  • Solid Python programming skills.
  • Solid math foundations: linear algebra, linear & non-linear optimization.
  • Strong scientific method; experience discovering innovations through experiments.
  • BONUS: Experience building deep neural net approaches for 3D scene perception.
  • BONUS: Expertise with indoor architectural layout estimation methods.
  • Must be creative, entrepreneurial, excited to push forward the state of the art.


  • Unique opportunity to deliver first-to-market, scene perception advancements to hundreds of millions of people.
  • Opportunity to publish novel research, participate in conferences, and collaborate with partners in academia.
  • We will sponsor U.S. work visas for international candidates.
  • We are in a great location in downtown Mountain View, California, one block from Caltrain & restaurants, and close to the Stevens Creek bicycle trail. We plan to open a Seattle office in 2021.
  • Location is flexible - experienced candidates may work from any location, as long as you have reliable internet access and can work most hours of a Pacific Time work day.
Apply for the job