Lead Machine Learning Engineer - 3D Data

NavVis
München, Germany
16 days ago

Role details

Contract type
Permanent contract
Employment type
Full-time (> 32 hours)
Experience level
Expert
Experience required
5 years minimum
Working hours
Regular working hours
Job source

Tech stack

Clean Code Principles 3d Models Artificial Intelligence Artificial Neural Networks Computer Vision C++ (Programming Language) Nvidia CUDA Computer Graphics Linux Python (Programming Language) Machine Learning OpenCL
+14 more
Tensorflow Computational Geometry Pytorch Deep Learning Gpu Programming Gaussian Scikit Learn Information Technology Slurm Machine Learning Operations Multiaccess Edge Computing Stable Diffusion Data Pipelines Databricks

Job description

At NavVis, we build spatial intelligence technology that depends on high-quality 3D data. This is a machine learning engineering role focused on 3D geometric data. We’re looking for an experienced engineer to develop and deploy machine learning models that operate on point clouds and 3D reconstructions, while raising the team’s expertise in applying ML and DL to 3D data and broadening how the team uses it.

You’ll work closely with cross-functional teams, including AI researchers, software engineers, and product designers, to build and ship models for 3D point cloud processing, filtering, surface reconstruction, and computational geometry. Most of your time goes into moving these methods from research into production, not just exploring them., * Design and optimize machine learning and deep learning models for 3D point cloud data, covering filtering, denoising, registration, and surface reconstruction

  • Own 3D data pipelines end-to-end, from preprocessing and augmentation to model adaptation and evaluation against ground-truth geometry
  • Integrate 3D ML solutions into existing workflows and products by collaborating with other teams
  • Adapt state-of-the-art 3D deep learning methods for production use, for example point-based architectures such as PointNet++, sparse voxel or convolutional networks, graph neural networks, and implicit or neural representations
  • Build scalable, reproducible training pipelines for large 3D datasets
  • Improve model performance by analyzing results against geometric and perceptual quality metrics

Requirements

If you have practical experience and a strong theoretical grounding in cutting-edge machine learning and deep learning techniques, and want to apply them to 3D geometric data, you’ll have the scope to shape how NavVis approaches 3D machine learning., * MS or PhD in computer vision, machine learning, computer science, or a related field, or equivalent practical experience

  • 5-7+ years of hands-on experience applying machine learning and deep learning to point cloud processing, filtering, registration, or surface reconstruction, with a strong theoretical grounding in the underlying methods
  • Familiarity with point cloud-specific deep learning architectures, for example PointNet++, sparse convolutional networks, graph neural networks, or transformer- and diffusion-based 3D models
  • Strong Python skills for writing efficient, maintainable code, and hands-on experience with libraries such as PyTorch, TensorFlow, and Scikit-learn
  • Solid foundation in the mathematics of geometry, linear algebra, and optimization
  • Fluent English communication skills, needed to collaborate with AI researchers, software engineers, and product designers, * Experience with computer graphics and GPU programming (CUDA, ROCm, or OpenCL)
  • Experience with neural 3D representations such as NeRF, Gaussian splatting, or signed distance functions
  • Familiarity with ML experimentation and orchestration tooling, for example MLflow, Ray, Databricks, Slurm, or Weights and Biases
  • Published research papers or open-sourced projects in a relevant ML/DL field
  • Experience with C++ and 3D libraries such as Open3D, PyTorch3D, PCL, or CGAL
  • Experience with Linux

Benefits & conditions

Pulled from the full job description Referral program Flexible schedule, * It’s important to take a break from work! We offer 30 days of paid time off per year

  • Affordable access to a vast network of fitness and wellness facilities through EGYM Wellpass subsidy
  • Deutschlandticket subsidy to support sustainable travel using public transport
  • We offer flexible working hours and a hybrid work setup, enabling you to plan your work around your life, and not your life around work!
  • We offer full visa and relocation support for international candidates
  • An attractive bike leasing model through JobRad, in line with our commitment towards sustainable mobility
  • A competitive compensation package that values the skills and experience you bring
  • Up to 4000 EUR employee referral bonus
  • Financial support for local language classes to help you in your journey of integrating into the culture!

About the company

When people use our technology to “bridge the gap” between the physical and digital worlds, they don’t just capture reality - they create a new one. In this new reality, they are smarter, more productive, more streamlined, and more creative - because they have the digital foundation to build the world they want to live in.

That’s what NavVis offers in all our products and services: the tools to not just map the world as it is, but to pave the way to a better future. To forge something new. Physical or digital, there is only one reality. And it’s the reality NavVis empowers people to build better., NavVis is a technologically-focused, global leader in reality capture and digital factory solutions. Prominent manufacturers and laser scanning professionals around the world trust our ground-breaking technology, designed to capture and share the built environment as photorealistic digital twins. By bridging the physical and digital worlds, we’re helping organizations make smarter decisions and build more efficient, connected operations. With over 300 people from around 60 nationalities and offices around the globe, we’re proud to be a truly international and diverse place to work. As a scaling company, we bring together the agility, innovation, and entrepreneurial mindset of a startup with the professionalism and reliability of an established enterprise - offering our team the best of both worlds. You’ll be part of an open culture that values trust, collaboration, and transparency. We are proud of our environment where diverse talent can thrive and where feedback fosters individual growth and development. Do you share our passion for disruptive technology and want to be part of our dynamic journey? At NavVis, you’ll own your projects, bring innovative ideas to life, and have real opportunities to grow your career and be recognized for your impact. We derive our strength from our diversity.

NavVis’ unwavering commitment to fostering an inclusive and diverse workplace has laid the foundation for our incredible growth. We thrive on the collective strength of our people who come from diverse backgrounds. We respect and value every experience associated with race, gender identity, sexual orientation, nationality, religion and disability. We do not discriminate on the basis of any of these, or other identities, and strongly encourage everyone to apply.

Together with you, we build NavVis!

Apply for this position

This job is hosted externally. Click below to view the full posting and apply.

Apply on de.indeed.com

Good distractions

Talks and stories from around this role — technically off-topic, practically not.

1:42 min

Navigating emerging hardware standardization in vendor programming ecosystems

Paul Graham Paul Graham · LIVE

52 sec

Running persistent Linux environments directly on Windows

Ben Breard Ben Breard · WWC 2025

2:22 min

Infrastructure barriers and compliance risks in research

Jeremy Murray Jeremy Murray · WWC Europe 2026

3:14 min

Structuring career paths and localized data architectures

Ulrich Wurstbauer +1 · LIVE

2:39 min

Experiencing core Linux capabilities for DevOps administration

Michael Cade · LIVE

5:20 min

Building neural search applications across diverse data modalities

Han Xiao · WWC 2022

Videos

See all

Related articles

See all