Industrial PhD - Graph Neural Networks

The situation:

MapsIndoors is a market leading platform for indoor wayfinding, and MapsPeople are working with some of the largest companies in the world. As a data visualization platform, we need to continue our automation efforts to reach our goal of producing "Maps in Minutes”. To obtain this, we invest heavily in map automation and together with Aalborg University and Innovation Fund Denmark we now have an open industrial PhD position.

As an Industrial PhD student, you will have equal hours at MapsPeople and at Aalborg University. However, all your research will relate to the project “Graph Neural Networks for Automated Map Generation”. At MapsPeople, you will join our automation team that works on applying Machine Learning to our work processes. At Aalborg University you will be supervised by professor Thomas Moeslund and become a member of the strong computer vision group: Visual Analysis and Perception lab:

The industrial PhD position is scheduled for three years.

Graph Neural Networks for Automated Map Generation:

You will work closely with our CIO, Data Automation Engineers, Deep Learning Engineers and Geo Data Specialists to research and develop novel graph neural network (GNN) algorithms that can automatically analyze less structured data including technical drawings. The prime goal for MapsPeople is to automate our production line to deliver and maintain MapsIndoors solutions fast and in a high and consistent quality.

We have well documented manual processes to bring CAD drawings to life in the MapsIndoors application and a deep understanding of the geodata domain. You bring the deep understanding of the deep learning domain so we together can reach the goal.

Specifically, your job is to...

  • Research and develop state-of-the-art performing GNN algorithms.
  • Be on the frontier of knowledge for MapsPeople, thus increase the knowledge and competency level in Neural Networks in general and GNN in particular.
  • Center your research around the workflow in MapsPeople and help solve classification and localization problems by applying supervised, unsupervised and self supervised geometric deep learning.
  • Establish knowledge and algorithms to automate and optimize data annotation and processing of the in-house vector data.
  • Develop an end-to-end mapping from detailed CAD drawings to simplified indoor and outdoor maps.
    • How can the annotation process be optimized?
    • How can significantly novel inputs be identified automatically?
    • How can objects, i.e., rooms, doors etc. be detected using GNN?
  • Work with the team i.e., Deployment, DevOps, Data and Machine Learning engineers to assist in bridging the gap between research and operations

Your profile:

To be successful in this position you hold a M.Sc. in Computer Science or related field and we expect you to have an interest in becoming an expert within or have experience within:

  • Machine learning algorithms and theory
  • (Applied) Deep learning, Graph neural networks, autoencoders, GANs and computer vision.
  • Data Science.
  • Optimization

If you have previous experience with geometric deep learning it is clearly an advantage.

On a personal level you are…

  • Creative and analytical.
  • Communicative with the ability to respectfully engage in discussions with people of different backgrounds.
  • Self-motivated with attention to detail and a drive to produce high quality work.
  • Team oriented.

You communicate fluently in English, and proficiency in Danish is an additional plus.

Your expectations:

You can expect a job with a lot of exciting professional and personal challenges. Your effort will make a difference for clients around the world, where our platform plays a central role in regards to indoor mapping, indoor positioning, and in the integration with various software and hardware technologies.

You can expect us to work together and pull together as a team.

You will experience modern office facilities with room for contemplation and an informal and dynamic company culture.

We offer a compensation package that matches your qualifications as well as flexitime, lunch, and retirement and healthcare plans.

MapsPeople - guiding you anywhere - anytime:

MapsPeople is a world-leading provider of indoor navigation through our flagship product ‘MapsIndoors’. For more than a decade, we’ve been a trusted partner and provider of Google Maps licencing products.

From our offices in Aalborg (HQ), Copenhagen, and Austin (USA), we strive to bring value to our customers and end-users. We are proud to serve some of the world’s most recognized companies and institutions, e.g. JPMorgan Chase & Co., Kohl's, KPMG, Google, LinkedIn, UEFA, Lego, City University of London, Copenhagen Airport, and many more.

We have a strong collaboration with Google Maps, and for several years, we have been a Google Maps Premier Partner.

At MapsPeople we value professionalism, creativity, and speed in an informal environment. 
We are 65+ dedicated employees who represent 16 different nationalities. To support our continued growth in 2020, we are recruiting a lot of new talent.


We look forward to receiving your resumé for this position. Please upload your documents in pdf-format by clicking on the 'Apply for this job'-button below. 

For more information, reach out to HR Business Partner, Mark Hellerup Knudsen on phone: +45 2227 2629 or

Navigate to MapsPeople:

Click to start navigation:

MapsPeople pin


Public Transport

NOTE for candidates: To comply with European legislation regarding the processing of personal information (GDPR), we only accept applications and resumes via the MapsPeople career site ( This means that only applications/resumes submitted via the career site are considered in the recruitment process. Applications and resumes sent to MapsPeople employees on e-mail are discarded without notice and without notification.
Be aware that some job sites that crawl our career site offer a "quick apply" button which in turn sends an email to us. Such e-mails are discarded and you will not receive a confirmation.

NOTE for headhunters/recruiting agencies: MapsPeople does not accept forwarding of any candidate resumes or candidate description from third parties including recruiting agencies or consultancy companies without an explicit and written agreement.
In case MapsPeople hires a candidate who has connected to MapsPeople or applied for a specific job via our career site ( or by other means, MapsPeople cannot be held accountable as well as no fee or compensation what-so ever will be paid to any recruiting agency even if this recruiting agency has forwarded a resume for that same candidate without explicit and written agreement with MapsPeople.

Or, know someone who would be a perfect fit? Let them know!

Workplace & Culture

We have the privilege of working daily with talented, passionate, and diverse colleagues. We enjoy working as a team and helping each other across departments as we see diversity to be our strength. It’s our sense of open and including communication that makes our workplace fun, innovative, and dynamic.

MapsPeople culture is based on trust, and it's a great blend of dedication to tasks and humour. Our culture is professional and goal-oriented while also having room for fun. When we achieve our goals, we celebrate! With diverse social events, we keep our culture strong.

MapsPeople culture can be seen in everything we do, it guides us anywhere anytime.

Already working at MapsPeople?

Let’s recruit together and find your next colleague.

  • Angee Clover Austin
  • Christian Müggler
  • Daniel Nielsen
  • Emil Siegenfeldt Nielsen
  • Fie Hovalt Christoffersen
  • Frederik Østerby Hansen
  • Gary Wood
  • Jane Haugaard Poulsen
  • Jannik Brouwers
  • Janni Rasmussen

Applicant tracking system by Teamtailor