
High-Performance Object Detection on Edge AI Hardware
A reported benchmark of more than 100 FPS using YOLOv8n on a Blaize X1600 GSP at 416 × 244 pixels.
Engineering work on edge hardware, video analytics architecture, model deployment and computer vision products. Each page says what was done and what is not claimed.

A reported benchmark of more than 100 FPS using YOLOv8n on a Blaize X1600 GSP at 416 × 244 pixels.

Engineering a video analytics architecture that supports NVIDIA and Blaize accelerators, with common services separated from hardware-specific inference.

An engineering workflow from trained model to deployable inference pipeline: assessment, conversion, optimisation, video integration and evaluation.

A product in development: computer vision on mobile video to flag visible road issues for review by inspection teams.
Tell us about your model, hardware or site, and we will tell you what we would evaluate first.
A site assessment not a sales pitch scoped before any commitment.