Senior Online Calibration & SLAM Engineer - On-Site

Senior Online Calibration & SLAM Engineer

Location: Toronto GTA or New York (on-site, periodic travel to hardware site) 

Type: Full-time 

Openings: 1 

Salary: CA$200K – $230K + equity

The role

Own how our client's capture devices stay calibrated after they leave the factory. Cameras and IMUs drift and mounts warp, so a device that shipped in spec can fall out of it. You'll build the systems that catch and correct that drift across the whole fleet, on solid SLAM/VIO foundations. This is a production role, not pure research, working with the computer vision and ML team, factory calibration and hardware engineering.

What you'll do

• Build continuous, in-pipeline camera calibration that refines extrinsics from hardware-synced video and IMU data, so fixes ship without device updates
• Harden the video-based calibration refinement system fleet-wide, including the monocular wrist-camera path
• Build fleet-wide drift monitoring and metrics, flag devices for recall, and own the calibration database and version history
• Maintain and improve the SLAM/VIO stack and have it report calibration error as a first-class output
• Coordinate with factory calibration and an overseas hardware team on intrinsics, IMU noise and bias, time sync and validation runs
• Ship interactive tools (Rerun, Gradio) to visualize trajectories, drift, reprojection error and per-device metrics

What you bring

• 5+ years building, maintaining or improving SLAM, visual-inertial odometry and Structure-from-Motion pipelines, including production systems
• Hands-on online or continuous calibration and rigorous camera + IMU calibration: extrinsics, intrinsics, IMU noise and bias, lens distortion, temporal sync
• A fleet and data-quality mindset across many noisy real-world devices
• Strong multi-view geometry, linear algebra and optimization fundamentals
• Clean, efficient, scalable C++ and Python
• Clear communication of calibration requirements to hardware teams across time zones

Nice to have

• Markerless or video-based calibration
• Kalibr, GTSAM or Ceres Solver; OpenCV, COLMAP, PyTorch, FFmpeg
• Fleet observability, calibration databases and versioning
• Rerun, Gradio or trajectory and dataset browsers; data pipelines at scale