PeopleofSuperintelligence

Researcher

Peter Kontschieder

Peter Kontschieder is a computer vision researcher whose work runs from decision forests and street-level scene understanding to neural 3D reconstruction. After a PhD at Graz University of Technology in 2013, he was a postdoctoral researcher in the Machine Intelligence and Perception group at Microsoft Research in Cambridge, UK, until 2016, and he received the 2015 Marr Prize for Deep Neural Decision Forests, which unifies classification trees with deep convolutional networks. In 2016 he founded Mapillary Research, coauthored the Mapillary Vistas Dataset in 2017, and joined Facebook in Zurich with his team after the 2020 acquisition of Mapillary. VR-NeRF in 2023 and MapAnything in 2025 are among his later coauthored works. He is a Director of Research Science at Meta, where he leads Meta Reality Labs' efforts in foundational Spatial AI and World Model development.

Now at
Meta
Badges

Timeline

  1. 2013–2016

    Graz doctorate, Microsoft postdoc

    He earns his PhD at Graz University of Technology in 2013. From 2013 to 2016, he is a postdoctoral researcher in the Machine Intelligence and Perception group at Microsoft Research in Cambridge, UK.

  2. 2015

    A Marr Prize for decision forests

    With Madalina Fiterau, Antonio Criminisi and Samuel Rota Bulo, he coauthors Deep Neural Decision Forests at ICCV 2015, a model that unifies classification trees with the representation learning of deep convolutional networks. Integrated into GoogLeNet, it reaches top-5 errors of 7.84% for one model and 6.38% for seven models on ImageNet validation data. He receives the 2015 Marr Prize for the work.

  3. 2016–2018

    Founding Mapillary Research

    In 2016, he joins Mapillary and founds Mapillary Research, its group for applied computer vision and machine learning. He heads a new research lab in Graz, Austria, with Gerhard Neuhold, and in 2018, as head of research at Mapillary, he talks on the Software Engineering Daily podcast about the research behind Mapillary's API for information about the physical world.

  4. 2017

    The Mapillary Vistas Dataset

    In 2017, he coauthors The Mapillary Vistas Dataset for Semantic Understanding of Street Scenes with Gerhard Neuhold, Tobias Ollmann and Samuel Rota Bulo at ICCV. The dataset holds 25,000 high-resolution street-level images annotated into 66 object categories, with instance-specific labels for 37 classes.

  5. 2020

    Joining Facebook in Zurich

    With the acquisition of Mapillary in 2020, he and his team join Facebook in Zurich.

  6. 2023

    VR-NeRF for walkable spaces

    In 2023, he coauthors VR-NeRF, presented at SIGGRAPH Asia 2023. The system captures walkable spaces, reconstructs them with neural radiance fields, and renders them in real time in virtual reality.

  7. 2025

    MapAnything 3D reconstruction model

    He is the final author of MapAnything, posted in September 2025: a transformer-based feed-forward model that takes one or more images, with optional inputs such as camera intrinsics, poses or depth, and directly regresses metric 3D scene geometry and cameras.

  8. As of October 2026

    Leading spatial AI at Meta

    He is a Director of Research Science at Meta and part of the Zurich site leadership. He leads Meta Reality Labs' efforts in foundational Spatial AI and World Model development.

Career

  1. Since 2020MetaDirector of Research Science
  2. From 2016MapillaryHead, Founder
  3. 2013–2016Microsoft ResearchPostdoctoral Researcher

Education

  1. 2013Graz University of TechnologyPhD

From the biography.

  1. 11% of AI doctorate holders from 2010 to 2016 founded or led a company or lab by 10 years after graduationIncludes Peter Kontschieder. We estimate how often AI doctorate holders founded or led a company or lab by 10 years after graduation, how waiting times compare between founders and top jobs, and which universities trained the most.
  2. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Peter Kontschieder. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.
  3. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes Meta. We estimate where the people shaping AI worked when their most-cited papers appeared: the company share was 31% in 2020–2024 and 8% in 1970–1974, and Google led all organizations in 2020–2024.
  4. Companies supply the most prior-work hires at 7 of the 9 frontier AI labsIncludes Meta. Our estimates of the organizations where 550 people worked directly before joining 9 frontier AI labs across 620 career transitions, comparing the main feeder organizations per lab.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Peter Kontschieder yet. Send us a free-licensed photo.

  1. Ingredients for Mapping the Metaverse – ETH Future Computing Laboratory | ETH Zurich efcl.ethz.ch. Checked 22 October 2026.
  2. Beyond Digital Twins: The Age of Visual World Models — Peter Kontschieder (Meta Reality AI Research) · Luma luma.com. Checked 22 October 2026.
  3. Deep Neural Decision Forests – ICCV 2015 Open Access Repository openaccess.thecvf.com. Checked 22 October 2026.
  4. Photo crowd-sourcing company Mapillary opens computer vision laboratory in Austria | Vision Systems Design vision-systems.com. Checked 22 October 2026.
  5. Computer Vision with Peter Kontschieder - Software Engineering Daily softwareengineeringdaily.com. Checked 22 October 2026.
  6. The Mapillary Vistas Dataset for Semantic Understanding of Street Scenes – ICCV 2017 Open Access Repository openaccess.thecvf.com. Checked 22 October 2026.
  7. [2311.02542] VR-NeRF: High-Fidelity Virtualized Walkable Spaces arxiv.org. Checked 22 October 2026.
  8. [2509.13414] MapAnything: Universal Feed-Forward Metric 3D Reconstruction arxiv.org. Checked 22 October 2026.