Bachelor/ Master Thesis - Deep Learning (f/m/d)

About us


It is sometimes hectic, very exciting, and always fun! Every day, you will find yourself facing different challenges, learning new things, working on cutting-edge technology, and building cool, high tech products.
Are you excited to build things that matter within an organization that’s pushing the boundaries of 3D mapping and computer vision? We have some of the world's leading experts in computer vision, image processing, and 3D mapping, and we work side by side with academia to advance these fields. Join the team that’s making it all happen.
At the end of the day, you’ll have a great sense of accomplishment, having contributed to building something completely innovative at the frontier of today's technology.


  • Take the chance to build & influence the performance of our high tech products instead of “only” running analyses or of doing “plain” software development
  • Use top notch IT equipment, including Bose headphones
  • Work closely with passionate, international & highly talented colleagues in a casual environment with flexible working hours
  • Find us in the heart of Munich, close to Stiglmaierplatz with great cafes, shops and subway around
  • Join your NavVis colleagues and continuously grow your biz/tech skills with courses from our internal “NavVis University” which brings you topics from expert internal & external trainers
  • Strong teams are built on strong bonds. Join us for bi-monthly BrownBag lunches, regular team-building activities, self-organized employee activities (sky is the limit on these!) and annual team off-sites
  • Take care of your well-being at work and enjoy our fully stocked kitchen with an original Italian coffee machine & all kinds of cereals and fruits


NavVis bridges the gap between the physical and digital world. With market-leading products like NavVis M6, NavVis VLX, and NavVis IndoorViewer, we empower our customers to capture and share the built environment as photorealistic digital twins.

Operating from offices in Munich, New York, and Shanghai, we employ over 200 people from more than 40 countries. Our company culture comprises people from all walks of life, including technologists, photographers, yoga practitioners, video gamers, runners, musicians, and more.

The Mapping and Perception team develops the software components for and around NavVis’ mobile mapping systems that enable simple, fast and large-scale mapping. A key focus is on building an accurate, realistic and complete reconstruction of the mapped environment that makes its representation useful for many applications.

Join us today!  For more information, visit

You will work on:

  • Literature research on pointcloud noise suppression and surface reconstruction.
  • Reproducing deep learning based pointcloud noise suppression and surface reconstruction results from literature.
  • Adapting approaches to be more appropriate to room scale scans from mobile scanners, since most published approaches focus on isolated objects of limited size.
  • Improving models and approaches to produce more accurate results on data from mobile scanners. (Optional for Bachelors thesis)
  • Making sure inference is capable of running on large scale pointclouds in acceptable time. (Optional for Bachelors thesis)
  • Creating training / evaluation data from simulation and real reference scans.
  • Training and validation of improved models on simulated and real data.
  • Comparing and evaluating results from classical and deep learning based approaches.

You have:

  • Done successful coursework in deep learning
  • Experience in Python specifically Numpy and Pytorch or Tensorflow
  • Excellent problem-solving skills
  • An independent, structured, and results-driven way of working
  • Interest in 3d reconstruction and mobile mapping systems

You are:

  • A student of computer science or related field (Electrical and Computer Engineering, Computational Science or similar)
  • Motivated to push the boundaries of large scale 3d reconstruction

We will be thrilled if you have:

  • Experience with C++ especially using Tensorflow and PCL
  • Experience with ROS and Linux
  • Implemented a deep learning project before
  • A clean coding style and experience using TDD
  • Knowledge of classical SLAM, calibration, pointcloud filtering, image processing, lidar data processing or sensor fusion techniques
  • Good contact to a professor or PhD student that may be willing to supervise you from the academic side (We can help finding a supervisor if you do not)