R&D Engineer – 3D Vision & Scene Reconstruction
Location: Berlin
Full time (40h/week)
Vision being the dominant human sense, eye tracking constitutes a powerful approach for understanding the human mind! At Pupil Labs, our mission is to provide cutting-edge eye-tracking solutions, which are more robust, accurate, accessible, and user-friendly than ever before. Already today, our products empower thousands of users in academia and industry, clinical surgeons, elite athletes, astronauts on the International Space Station, and many more. Unlocking the full potential of eye-tracking technology increasingly relies on 3D scene understanding — from pose and depth to full environment reconstruction.
To support our efforts, we are looking to grow our Berlin-based team with a full-time R&D Engineer – 3D Vision & Scene Reconstruction, with experience in multi-view geometry, 3D representations, and modern rendering techniques such as Gaussian Splatting. This is an on-site position (with up to two home-office days per week).
Pupil Labs offers a competitive salary, flexible work arrangements, a great team of coworkers, a young and dynamic company structure, and a culture of participation and feedback.
You are excited about joining an ambitious, international, diverse, interdisciplinary, young, enthusiastic, and talented team of researchers and engineers? You have a growth mindset, thrive in fast-paced work environments, and enjoy building real systems with geometric depth? Then we are looking forward to hearing from you!
What you would do
Design and implement 3D reconstruction pipelines based on multi-view geometry, depth estimation, and point cloud fusion.
Explore and integrate neural and non-neural 3D representations, including Gaussian Splatting, NeRFs, voxel grids, and mesh-based techniques.
Work with large-scale image/video data to reconstruct detailed, accurate scenes, and assist with aligning 3D reconstructions to real-world coordinate systems.
Collaborate with researchers and engineers to evaluate, optimize, and deploy geometry-aware models into real products.
Help translate cutting-edge ideas in structure-from-motion, SLAM, and differentiable rendering into robust, maintainable code.
Who you are
You have a Master’s or PhD in computer science, applied mathematics, computer vision, graphics, or a related field, or equivalent practical experience.
You have strong foundations in multi-view geometry, camera modeling, depth estimation, and 3D scene reconstruction.
You have worked with tools like COLMAP, Open3D, OpenCV, or similar, and are comfortable building pipelines around them.
You are familiar with modern scene representations, such as Gaussian Splatting, NeRFs, TSDFs, meshes, or point clouds.
You are proficient in Python and PyTorch.
You care about software quality, writing maintainable, well-documented, and testable code.
Ideally, you have experience with OpenGL or WebGL, and understand how 3D representations interface with real-time rendering.
You are comfortable in written and spoken English.
Perks
A beautiful office in the heart of Berlin.
Up to two home-office days per week.
15 additional mobile-office days per year.
Continued learning and professional development (we will sponsor you to attend relevant scientific/developer conferences.
Flexible working hours.
Publishing of scientific articles.
6 weeks of holidays per year.
Apply
Please submit your application here.
Vision being the dominant human sense, eye tracking constitutes a powerful approach for understanding the human mind! At Pupil Labs, our mission is to provide cutting-edge eye-tracking solutions, which are more robust, accurate, accessible, and user-friendly than ever before. Already today, our products empower thousands of users in academia and industry, clinical surgeons, elite athletes, astronauts on the International Space Station, and many more. Unlocking the full potential of eye-tracking technology increasingly relies on 3D scene understanding — from pose and depth to full environment reconstruction.
To support our efforts, we are looking to grow our Berlin-based team with a full-time R&D Engineer – 3D Vision & Scene Reconstruction, with experience in multi-view geometry, 3D representations, and modern rendering techniques such as Gaussian Splatting. This is an on-site position (with up to two home-office days per week).
Pupil Labs offers a competitive salary, flexible work arrangements, a great team of coworkers, a young and dynamic company structure, and a culture of participation and feedback.
You are excited about joining an ambitious, international, diverse, interdisciplinary, young, enthusiastic, and talented team of researchers and engineers? You have a growth mindset, thrive in fast-paced work environments, and enjoy building real systems with geometric depth? Then we are looking forward to hearing from you!
What you would do
Design and implement 3D reconstruction pipelines based on multi-view geometry, depth estimation, and point cloud fusion.
Explore and integrate neural and non-neural 3D representations, including Gaussian Splatting, NeRFs, voxel grids, and mesh-based techniques.
Work with large-scale image/video data to reconstruct detailed, accurate scenes, and assist with aligning 3D reconstructions to real-world coordinate systems.
Collaborate with researchers and engineers to evaluate, optimize, and deploy geometry-aware models into real products.
Help translate cutting-edge ideas in structure-from-motion, SLAM, and differentiable rendering into robust, maintainable code.
Who you are
You have a Master’s or PhD in computer science, applied mathematics, computer vision, graphics, or a related field, or equivalent practical experience.
You have strong foundations in multi-view geometry, camera modeling, depth estimation, and 3D scene reconstruction.
You have worked with tools like COLMAP, Open3D, OpenCV, or similar, and are comfortable building pipelines around them.
You are familiar with modern scene representations, such as Gaussian Splatting, NeRFs, TSDFs, meshes, or point clouds.
You are proficient in Python and PyTorch.
You care about software quality, writing maintainable, well-documented, and testable code.
Ideally, you have experience with OpenGL or WebGL, and understand how 3D representations interface with real-time rendering.
You are comfortable in written and spoken English.
Perks
A beautiful office in the heart of Berlin.
Up to two home-office days per week.
15 additional mobile-office days per year.
Continued learning and professional development (we will sponsor you to attend relevant scientific/developer conferences.
Flexible working hours.
Publishing of scientific articles.
6 weeks of holidays per year.
Apply
Please submit your application here.
Vision being the dominant human sense, eye tracking constitutes a powerful approach for understanding the human mind! At Pupil Labs, our mission is to provide cutting-edge eye-tracking solutions, which are more robust, accurate, accessible, and user-friendly than ever before. Already today, our products empower thousands of users in academia and industry, clinical surgeons, elite athletes, astronauts on the International Space Station, and many more. Unlocking the full potential of eye-tracking technology increasingly relies on 3D scene understanding — from pose and depth to full environment reconstruction.
To support our efforts, we are looking to grow our Berlin-based team with a full-time R&D Engineer – 3D Vision & Scene Reconstruction, with experience in multi-view geometry, 3D representations, and modern rendering techniques such as Gaussian Splatting. This is an on-site position (with up to two home-office days per week).
Pupil Labs offers a competitive salary, flexible work arrangements, a great team of coworkers, a young and dynamic company structure, and a culture of participation and feedback.
You are excited about joining an ambitious, international, diverse, interdisciplinary, young, enthusiastic, and talented team of researchers and engineers? You have a growth mindset, thrive in fast-paced work environments, and enjoy building real systems with geometric depth? Then we are looking forward to hearing from you!
What you would do
Design and implement 3D reconstruction pipelines based on multi-view geometry, depth estimation, and point cloud fusion.
Explore and integrate neural and non-neural 3D representations, including Gaussian Splatting, NeRFs, voxel grids, and mesh-based techniques.
Work with large-scale image/video data to reconstruct detailed, accurate scenes, and assist with aligning 3D reconstructions to real-world coordinate systems.
Collaborate with researchers and engineers to evaluate, optimize, and deploy geometry-aware models into real products.
Help translate cutting-edge ideas in structure-from-motion, SLAM, and differentiable rendering into robust, maintainable code.
Who you are
You have a Master’s or PhD in computer science, applied mathematics, computer vision, graphics, or a related field, or equivalent practical experience.
You have strong foundations in multi-view geometry, camera modeling, depth estimation, and 3D scene reconstruction.
You have worked with tools like COLMAP, Open3D, OpenCV, or similar, and are comfortable building pipelines around them.
You are familiar with modern scene representations, such as Gaussian Splatting, NeRFs, TSDFs, meshes, or point clouds.
You are proficient in Python and PyTorch.
You care about software quality, writing maintainable, well-documented, and testable code.
Ideally, you have experience with OpenGL or WebGL, and understand how 3D representations interface with real-time rendering.
You are comfortable in written and spoken English.
Perks
A beautiful office in the heart of Berlin.
Up to two home-office days per week.
15 additional mobile-office days per year.
Continued learning and professional development (we will sponsor you to attend relevant scientific/developer conferences.
Flexible working hours.
Publishing of scientific articles.
6 weeks of holidays per year.
Apply
Please submit your application here.