Agricultural Robotics Computer Vision Engineer

Role Summary Build and improve vision-based perception for agricultural robots, with strong emphasis on dataset curation, annotation strategy, training data quality, and rigorous model evaluation that translates to measurable field performance improvements.

What You Will Do

• Design and iterate on perception pipelines for crops, weeds, fruit, rows, and obstacles (detection, segmentation, tracking)
• Define labeling schemas and write clear annotation guidelines (boxes, polygons, instance masks, keypoints)
• Run QA evaluation to verify label consistency and ground-truth reliability across datasets
• Own dataset curation workflows: sampling strategy, class balance, long-tail coverage, and hard-negative mining
• Execute training/evaluation loops and analyze error modes using mAP, IoU, precision/recall, and confusion matrices
• Collaborate with robotics engineers on ROS2 integration, sensor fusion, and edge deployment constraints

Required Qualifications

• Mid-Senior experience delivering production-like computer vision or robotics perception systems
• Strong knowledge of object detection and semantic/instance segmentation
• Hands-on experience with dataset creation, labeling workflows, and training data quality processes
• Familiarity with robotics stacks and real-time deployment considerations (e.g., ROS2)

Workplace This is a Remote , FULL_TIME role based in the US.

Apply with a resume and relevant project links highlighting perception modeling, evaluation, dataset curation, and/or edge deployment.