Computer Vision Engineer

We are looking for a hands-on Computer Vision Engineer to work on the models that power the industrial AI platform. You will own computer vision problems end-to-end, from data strategy and annotation to model development, edge deployment, and production monitoring for some of the most complex vision problems in heavy industry.

Responsibilities:

• Own computer vision problems end-to-end across the industrial portfolio.
• Build models for hard vision challenges, including novel defect types, extreme class imbalance, multi-camera fusion, low-light/high-noise environments, and real-time inference on constrained edge hardware.
• Evaluate and adopt models including YOLOv26 SAM 3 Vision Transformers (DINOv2 Swin), Grounding DINO, RF-DETR, YOLO-World, and CLIP.
• Contribute to model training pipelines, DVC, CVAT, Roboflow, Label Studio, W& B, MLflow, TensorRT, ONNX, OpenVINO, and CI/CD.
• Drive inference optimization using quantization, pruning, knowledge distillation, and batching strategies.
• Debug production issues and improve deployed models.
• Champion a data-centric AI approach.
• Build evaluation frameworks and partner with product, field engineering, operations, and leadership.

Requirements:

• Bachelor's degree in computer science, AI/ML, electrical engineering, or a related field.
• 1-4 years of hands-on computer vision experience. Python, PyTorch, OpenCV, Albumentations, image/video processing.
• Experience in Object Detection, Instance/Semantic/Panoptic Segmentation, Anomaly Detection, Pose Estimation, and Tracking.
• Experience with YOLO, RT-DETR, DETR, RF-DETR, SAM/SAM2 ResNet, EfficientNet, ConvNeXt, and Vision Transformers.
• Production deployment using TensorRT, ONNX Runtime, or OpenVINO. Docker, Kubernetes, and AWS/Azure/GCP.
• Strong problem-solving, startup experience, and communication skills.

Good to Have:

• Industrial/manufacturing domain experience with PLCs/SCADA. Grounding DINO, YOLO-World, CLIP, DINOv2 SAM, 3Florence, GPT-4o Vision, Gemini, LLaVA, 3D Vision, depth estimation, point-cloud processing, and multi-camera calibration.
• ByteTrack, BoT-SORT, video analytics. Open-source contributions, CVPR/ECCV/ICCV/NeurIPS publications, or Kaggle results. Required Skills: Computer Vision, Python.