Production Computer Vision Engineer — Robotics & Imaging

CircuitHub, Inc.; Cambridge; On-site; GBP 80,000 - 125,000; Full time

Job summary

CircuitHub is reshaping electronics manufacturing with The Grid, a live robotics platform delivering high-mix, small-batch assembly at scale in the UK. The role focuses on advancing computer vision from prototype to production, leveraging Python/C++ and cutting‑edge libraries to improve inspection, part detection and alignment.

You will work with high-speed cameras, GPU pipelines and a data‑driven loop to drive real‑time decisions and measurable improvements in yield and throughput.

Qualifications

• Experience shipping computer vision systems into production.
• Strong Python or C++ software engineering skills.
• Mix of classical vision and deep learning; choose the right tool.
• Experience with industrial or machine-vision cameras is a bonus.
• Ownership of model development from data capture to deployment.

Responsibilities

• Inspect boards. Improve automated optical inspection for defects with AI assistance.
• Train and deploy models to locate parts in bins, center on nozzle, handle clumped or dropped picks.
• Perform fiducial detection, calibration and image alignment at sub-millimetre precision.
• Build the data loop: define data to capture, create labeling tools, use operator verdicts for training data.
• Utilize high-speed cameras and GPU pipelines to push vision toward real time.
• Track production accuracy, identify error sources, and validate changes against previous results.

Skills

Computer vision in production

Python

C++

Deep learning

Tools

OpenCV

PyTorch

ONNX Runtime

libtorch

Job description

CircuitHub is reshaping electronics manufacturing with The Grid, a live robotics platform delivering high-mix, small-batch assembly at scale in the UK. The role focuses on advancing computer vision from prototype to production, leveraging Python/C++ and cutting‑edge libraries to improve inspection, part detection and alignment.

You will work with high-speed cameras, GPU pipelines and a data‑driven loop to drive real‑time decisions and measurable improvements in yield and throughput.