Computer Vision Software Engineer

📋 Description

• • Own the full lifecycle of next-generation computer-vision systems that inspect every metal and plastic container we produce for the world’s largest food, beverage, and personal-care brands. From white-board concept to factory-floor deployment, you will architect, code, test, and tune algorithms that spot defects humans cannot see—at line speeds exceeding 2,000 parts per minute.
• • Design and implement high-performance C++ and Python modules that integrate tightly with motion-control PLCs, robotics, and high-speed cameras. You will squeeze every millisecond out of GPUs and edge devices to guarantee sub-100 ms inspection cycles while maintaining 99.9 % uptime across 24/7 manufacturing lines.
• • Build and continuously retrain deep-learning models using TensorFlow, PyTorch, and OpenCV to classify, segment, and measure micron-level anomalies on reflective, curved, and printed surfaces. You will own the MLOps pipeline—data versioning, experiment tracking, CI/CD, containerized deployment, and real-time model monitoring—so every improvement reaches production safely and quickly.
• • Install, calibrate, and commission vision systems on new and retrofitted production lines in Waukegan and at sister plants across North America. Expect hands-on days climbing catwalks, aligning cameras, and optimizing lighting, followed by nights analyzing terabytes of image data to prove statistical process capability.
• • Provide Tier-3 technical support to manufacturing and field-service teams. When a line goes down at 2 a.m. because a glare artifact slipped through, you will triage remotely, roll back models, or hop in the car to debug on-site—whichever gets safe food packaging back on the belt fastest.
• • Collaborate daily with mechanical, controls, and quality engineers to translate customer complaints and SPC trends into algorithmic solutions. You will present findings in gemba walks, Kaizen events, and quarterly business reviews, translating complex CV metrics into clear ROI and scrap-reduction numbers plant managers trust.
• • Extend our internal labeling toolset and front-end dashboards (Qt, React, or Streamlit) so operators can annotate images, adjust thresholds, and visualize defect heatmaps without a PhD. Your UX choices directly reduce training time and empower 1,000+ plant employees to interact with AI confidently.
• • Champion code quality and engineering rigor: write unit tests, enforce style guides, lead peer reviews, and mentor junior developers. You will establish SLAs, latency budgets, and automated regression suites that safeguard the $1 billion in annual product volume touched by your software.
• • Stay ahead of the curve—evaluate emerging 3-D sensors, hyperspectral cameras, and transformer-based architectures. Run quarterly tech spikes and publish internal white papers that influence Silgan’s five-year vision roadmap and patent portfolio.
• • Drive measurable impact: every 0.1 % false-reject reduction you achieve saves 2 million containers and 50 tons of metal per year, directly supporting Silgan’s sustainability pledge and our customers’ zero-waste goals.