The FactoryPulse demonstration on the Qualcomm Dragonwing IQ-9075 EVK implements three Vision AI use cases: Defect Detection, which analyzes a live conveyor belt video feed to detect faulty objects with bounding box overlays; Factory Digital Twin, which surfaces defect events with a red blinking alarm and a graph of faulty objects over time; and Worker Safety, which detects people and their Personal Protective Equipment (safety helmet and vest) in restricted factory areas. The demonstration includes 3 Vision AI models: Edge Impulse's YOLO Pro model for defect detection, and two TensorFlow Lite (TFLite) models for detecting protective gear and tracking people near restricted areas. All models are optimized for the Hexagon Tensor Processor (HTP) accelerators and run across the two HTP cores of the Dragonwing IQ-9075 EVK. The deployment combines the Qualcomm IM SDK, Edge Impulse trained models, and the Qt Framework (QML) to simplify Vision AI pipeline creation through a declarative approach. The Qualcomm IM SDK is a unified solution built on the open-source GStreamer framework, providing a catalog of plug-ins optimized for Qualcomm hardware, including the Adreno GPU, Spectra ISP, and Hexagon NPU. These hardware-accelerated plug-ins enable zero-copy buffer sharing, efficient high-resolution video streaming, direct data flow between camera, AI, and display components, and reduced CPU and memory overhead.