Senior Computer Vision Engineer, Real-Time 3D Tracking
About The Position
Fabstation puts AR work instructions directly onto structural steel during fabrication. Our software runs on AR tablets and headsets on real shop floors across North America, overlaying CAD-driven guidance onto multi-tonne steel assemblies during fit-up, welding, and inspection. We are cash-flow positive, growing fast, and used daily in production. We are hiring a Senior Computer Vision Engineer to own our real-time 3D object tracking capability: locking a CAD model onto a physical steel member and keeping it locked as the user moves around it. Tracking is core to our product, and we build it in-house. You will build on internal groundwork already underway and own the road from there: architecture, implementation, optimization, and production hardening. This is a role for someone who wants to own a hard, central problem end-to-end and see it shipped into the hands of real tradespeople, not maintain someone else's pipeline.
Requirements
- 5+ years of professional experience in computer vision or a closely related field, with a focus on real-time systems
- Shipped at least one real-time computer vision feature on a mobile platform (iOS or Android)
- Strong modern C++; Python is an asset
- Solid foundation in 3D geometry: camera models, calibration, projective geometry, and linear algebra
- Working command of nonlinear least-squares optimization (Gauss–Newton / Levenberg–Marquardt) and pose parameterization on SE(3)
- Hands-on experience with pose estimation and object tracking algorithms
- Experience integrating native C++ libraries with Unity, or equivalent native-mobile plugin experience
- Strong profiling and optimization skills on ARM (frame budgets, Instruments or equivalent)
- Working proficiency with OpenCV
Nice To Haves
- Model-based 3D object tracking (region-based, edge-based, or hybrid approaches)
- Metal and GPU compute on Apple silicon
- Visual-inertial odometry or SLAM; ARKit (ARCore an asset)
- State estimation and multi-sensor fusion: Kalman filtering, factor graphs (Ceres, g2o, GTSAM), extrinsic calibration, and time synchronization
- On-device ML deployment: Core ML, ONNX Runtime, or Unity Sentis; learning-based pose estimation or refinement that has shipped on-device
- CAD data handling: mesh processing and decimation, IFC/STEP-style exports
- Background in industrial vision, robotics, or augmented reality
Responsibilities
- Own the on-device object tracking core end-to-end, from architecture through production
- Design and implement real-time, model-based 6DoF pose estimation and continuous tracking for textureless steel structures, using CAD models as the tracking reference
- Fuse object tracking with platform world tracking (visual-inertial odometry), depth, and IMU data; own tracking-loss detection, drift management, and recovery
- Optimize native C++ vision pipelines for iOS and Apple silicon: rendering, CPU/GPU budgets, and sustained frame rates on iPad or headset hardware
- Extend tracking to large structures: partial visibility, multi-region strategies, and CAD mesh preprocessing from IFC/STEP - type exports
- Integrate native modules with Unity via a plugin bridge, and work with our Unity team from prototype through release
- Define tracking quality metrics and benchmark against real shop-floor conditions: occlusion, mixed lighting, and cluttered scenes of visually similar material
Benefits
- Cash-flow positive, growing fast, and used daily in production.
- Direct ownership
- Direct access to real customer shop floors
- Short path from your code to a tradesperson's hands
- Equal opportunity employer
Stand Out From the Crowd
Upload your resume and get instant feedback on how well it matches this job.
What This Job Offers
Job Type
Full-time
Career Level
Senior
Education Level
No Education Listed
