Software Engineer, Computer Vision

Role overview Develop computer vision systems for automated electronics assembly and inspection. The role combines classical image processing and machine learning, moving solutions from development into production across tasks such as checking circuit boards, guiding component placement, and detecting unreliable picks.

Responsibilities - Improve optical inspection that identifies missing, misplaced, rotated, or poorly soldered components. - Train and deploy models to locate parts in bins and trays, align them for pickup, and detect failed or clumped picks. - Build fiducial detection, calibration, and image alignment for precise component placement. - Decide what visual data to collect, create labeling tools, and convert operator feedback into training examples. - Work with high-speed cameras and GPU capture systems to reduce processing delays. - Track accuracy in production, investigate error patterns, and compare changes against historical results.

Requirements - Experience shipping computer vision systems into production. - Strong software engineering ability in Python, C++, or a comparable language. - Practical experience with both classical vision methods and deep learning, including judgment about when to use each. - Comfort working with physical equipment and real-world imaging factors such as lighting, lenses, and part orientation. - Ability to use coding agents responsibly and take ownership of the resulting work.

Nice to have - Experience with industrial or machine-vision cameras. - Relevant tools include OpenCV, PyTorch, ONNX Runtime, and GPU-based image processing; services use Python and Haskell, with PostgreSQL and Nix-based infrastructure.

Benefits and work setup - Compensation is listed as £80,000–£125,000 in the UK or $125,000–$250,000 in the US, depending on experience, plus equity.