General Information

Full Name Rahul Raguram


  • 2012
    PhD, Computer Science
    University of North Carolina at Chapel Hill
    • Advisors
      • Jan-Michael Frahm and Marc Pollefeys
    • Dissertation
      • Efficient Algorithms for Robust Estimation
      • Developed efficient algorithms for robust estimation, with a focus on problems in 3D computer vision. Worked on large-scale SfM pipelines, including some of the first systems for large / city-scale reconstruction from internet photo collections.
  • 2007
    MS, Electrical and Computer Engineering
    University of Arizona
    • Thesis
      • Improved Resolution Scalability for Bi-Level Image Data in JPEG2000
      • Analyzed the JPEG2000 algorithm with a focus on binary images, and developed some interesting tricks to improve visual / perceptual quality.


  • 2017-present
    Research Manager
    Apple, Inc., Cupertino, CA
    • Managing a team of engineers as part of the 3D Vision Team
    • We work on technology and algorithms to process large scale 3D datasets and build features that ship to millions of devices
    • Visual localization, slam, SfM, camera calibration, gnss, bundle adjustment, 3D reconstruction
  • 2012-2017
    Software Engineer, Senior Software Engineer
    Apple, Inc., Cupertino, CA
    • Worked on a number of large-scale 3D products, including 3D Flyover.
    • Bundle adjustment, calibration, slam, 3D reconstruction
  • 2011
    Intern, Lightfield/Geo Team
    Google, Inc., Seattle, WA
    • Designed and implemented a system for semantic segmentation of urban imagery and 3D point cloud data. The interesting component here was combining image appearance and geometry in a semi-supervised learning framework.
  • 2010
    Research Intern
    Los Alamos National Labs, Los Alamos, NM
    • Developed and implemented algorithms for 3D reconstruction and object tracking from high resolution overhead imagery.
  • 2010
    Visiting Researcher
    Center for Machine Perception, Czech Technical University, Prague
    • Worked on a next-generation robust estimation framework, in collaboration with researchers at CTU Prague. The system combined recent research efforts at UNC and CTU Prague.