senior-computer-vision

Production inference optimization uses quantization, TensorRT, and ONNX export to reduce model size and latency. This Skill covers these techniques across PyTorch, TensorFlow, and related frameworks to deploy vision models efficiently at scale.

Yes. This Skill diagnoses and constructs object detection and semantic/instance segmentation workflows using modern architectures in both PyTorch and TensorFlow, including training, inference optimization, and real-time processing.

Dataset pipelines require augmentation, labeling integration, and data quality checks before training. This Skill provides tooling to construct these pipelines end-to-end, ensuring reproducible experiment workflows and production-ready datasets.

This Skill works across PyTorch, TensorFlow, OpenCV, YOLO, SAM, and diffusion models, supporting object detection, segmentation, 3D vision, and end-to-end model training and deployment across multiple frameworks.

Yes. This Skill addresses real-time processing, production deployment, monitoring, and inference optimization for both image and video workflows, enabling end-to-end deployment of vision systems at scale.

This Skill delivers reproducible experiment pipelines with automated tooling, monitoring, and production deployment workflows, enabling consistent training, inference, and operational tracking of computer vision systems.