
Description
We're looking for a 3D Computer Vision Engineer — on a contract basis — to help shape software that turns raw 3D scan data into reliable, production-grade measurements and pass/fail decisions in real-world environments.
What you’ll do
Develop pipelines for:
Point cloud processing, filtering, and denoising
Scan-to-CAD alignment (ICP + feature-based methods)
Feature extraction and primitive fitting (planes, edges, holes, slots, chamfers, fillets)
Dimensional measurement, tolerance evaluation, and per-point deviation field computation
Translate raw geometry into clear inspection outputs (pass/fail, deviations, annotations)
Generate per-point deviation heatmaps for visual review in the 3D viewer
Handle calibration, drift, noise, and incomplete data
Ensure robustness across materials, lighting, and shop-floor conditions
Work directly with QC managers on the shop floor to validate and iterate
Ship working software on a short feedback loop — prototype, test in the field, harden
Support live customer deployments
Help define how UI/UX is optimized for operators
Contribute to end-to-end workflows, not just isolated algorithms
This is primarily a remote contract role, with approximately 20% of working time hands-on at a U.S. or European facility working directly with the 3D camera system to develop and test software against real hardware. Infrequent travel to client sites for system setup and validation is expected as needed.

Who You Are
You think in geometry, not just models, and care deeply about accuracy, robustness, and repeatability. You’re comfortable moving from research to production and don’t get stuck chasing perfect theory—you ship working systems, learn from real-world data, and iterate quickly.
Required Skills & Experience
Degree in Computer Science, Robotics, or related field
Strong experience in 3D computer vision / geometry processing
Proficiency in Python; C++ a strong plus
Experience shipping production-quality computer vision or measurement pipelines (not just research code)
Hands-on experience with:
Real sensor data (not just simulated or academic datasets)
Point cloud libraries (PCL, Open3D, or similar)
Registration (ICP, feature-based alignment)
Geometric primitive fitting (RANSAC, circle/cylinder/plane fitting)
Sensor calibration and depth systems
Nice to have
Inspection, metrology, or manufacturing systems experience
Parsing CAD models (DXF, STEP) and aligning scan data to CAD geometry
Deployment of systems into production environments
Robotics or industrial automation background
Why Toggle
Work on unsolved problems in the physical world and built environment
Define and build new systems used directly in factory and construction pre-fab production
Human centric with a priority on safety and usability
High ownership and fast iteration
Small, technical, global and collaborative team
Get curated roles, career notes, and practical perspective for moving in, out, or around architecture.
Occasional dispatches for designers exploring what comes next. No hard sell.