Staff Computer Vision Engineer - Localization
SE3 Labs
Staff Computer Vision Engineer - Localization
SE3 Labs
3D spatial AI SaaS for insights
Solve the hardest estimation and systems problems in our visual localization stack, from core algorithms to deployed software.
SE3 Labs builds spatial intelligence and autonomy software for unmanned systems. We combine 3D computer vision, state estimation and AI to help autonomous platforms perceive their surroundings, know where they are and operate when GPS and other sensors become unreliable. Defence is a core application of our work today.
We are growing to a team of 100+ people. Join at a stage where you can shape our localization technology, help build the team and grow your responsibilities as the company expands.
Our founding team’s research has received more than 90,000 citations. You will work closely with technical founders and engineers with deep experience in computer vision and autonomous systems, turning that research depth into systems that work in the field.
As a Staff Engineer, you will shape algorithm and architecture decisions, implement critical parts of the system, and help other engineers solve problems across sensors, estimation and onboard compute.
This is a hands-on individual contributor role. You will take ideas from research and recorded data through production code, live sensor testing and flight, working closely with the Team Lead and our robotics software, embedded, hardware and field teams.
VIO, SLAM and Sensor Fusion: Build and improve localization algorithms for autonomous systems operating with degraded or unavailable GNSS. Our initial focus is UAVs, with methods that can extend to other robotic platforms.
Robustness: Solve difficult failures caused by low light, rain, fog, motion blur, vibration and sensor degradation. Investigate better estimation, learned methods and data where each can improve measured performance.
Mapping and Relocalization: Develop mapping, loop closure, relocalization and map-based navigation. Make pose, velocity and uncertainty estimates useful to planning, control and the rest of the autonomy stack.
Calibration and Timing: Solve problems in camera and IMU calibration, sensor synchronization, rolling shutter, latency and frame drops. Work across algorithms and integration to find the root cause.
Real-Time Deployment: Design and optimize production software for Jetson-class platforms and other constrained CPUs and GPUs. Measure the tradeoffs between accuracy, robustness, latency, memory and power.
Evaluation: Build the evaluation and replay tools that let us reproduce failures, compare approaches and detect regressions. Validate improvements on recorded data and deployed systems.
Technical Leadership: Lead difficult technical projects, review designs and code, and teach other engineers what you learn. Help the team choose where deeper algorithm work will make the largest difference.
You have personally built and deployed VIO, SLAM, localization or sensor-fusion systems. You can explain the parts you implemented, the alternatives you considered, the failures you solved and the measured results.
You have solved problems that span algorithms, sensors and runtime constraints, and can show improvements in accuracy, robustness or compute cost.
You have strong foundations in 3D geometry, linear algebra, probability, estimation and optimization, with practical experience in methods such as EKFs, factor graphs, nonlinear least squares and bundle adjustment.
You write production C++ for Linux systems and use Python for analysis, evaluation and tooling. Rust experience is a plus.
You have worked with real cameras and IMUs, calibration and timing. You use logs, reproducible experiments and regression tests to establish why a change works.
You have led substantial technical work and helped other engineers improve their designs and implementation. You remain comfortable writing and debugging the critical code yourself.
You have a master’s degree, PhD or equivalent practical experience in a relevant technical field.
You speak fluent English; German is a plus. Given the nature of SE3’s work in the defence sector, candidates must be eligible to work on defence-related projects and, where required, obtain the relevant security clearance.
Experience deploying perception systems on Nvidia Jetson, embedded CPUs/GPUs, writing CUDA kernels, optimizing deep neural networks for edge deployment.
Experience with LWIR thermal cameras (for navigation), rolling shutter effects, visual-inertial calibration, multi-camera rigs, fisheye/wide-angle lenses or monocular depth.
Experience with GNSS-denied navigation in UAV context, terrain-relative navigation, map-based localization and re-localization.
Experience building evaluation datasets, replay systems, perception observability, log tooling, and automated regression pipelines.
Publications or open-source contributions in SLAM, VIO, 3D computer vision, state estimation, or robotics are welcome, but not required.
Real Ownership: Participate in SE3’s growth through our employee participation programme
Work Environment to Excel at: Find a high-quality and modern toolset, flat hierarchy and a skilled, driven team which will push the limits together with you
AI-boosted workplace: We empower every employee with best-in-class AI tools and agents, giving you unlimited access to the latest advancements and a true productivity edge
31 Paid Days Off: 30 vacation days plus your birthday off
Well-being Access: Free membership to gyms, yoga, and wellness through EGYM Wellpass
Team Offsites: We go on 2 company retreats per year, our last ones were in Italy and Greece
Mentorship: Work closely with experienced engineers and get paired with a professional mentor to support your growth
Intro Call (30 min): Let’s get to know each other
Technical Conversation (45 Mins): High level discussion about your technical experience and what we are working on
Coding / Algorithm Interview: We work through a practical perception, estimation, or robotics software problem together
Technical Interview (90 mins): We dive into your experience with perception, state estimation, SLAM/VIO, robotics, and real-world debugging
Team Dinner / Lunch: Meet the team and see whether it is a good fit on both sides
SE3 Labs builds an AI platform that turns unstructured 3D spatial data into actionable insights. It combines 3D computer vision with large language models to create a SaaS product called SpatialGPT, a “3D Vision as a Service” that lets users query and analyze 3D data using natural language. The system processes real-time object detection and semantic analysis even where GPS is unavailable, enabling workloads in defense, smart infrastructure, construction, robotics, and urban analytics. SE3 Labs differentiates itself through its highly experienced team from the Technical University of Munich and industry backgrounds, focus on dual-use applications and European sovereignty, and a subscription-based model with partnerships to tailor implementations for specific use cases. Its goal is to provide accessible, intelligent analysis of spatial data across civilian and defense sectors, helping organizations extract insights from complex 3D environments.
Explore more opportunities at SE3 Labs