Notes & Essays

Writing

Notes on shipping computer vision and generative models — quantization, runtime quirks, and the parts of deployment that papers leave out.

RSS feed
  1. Latest 14 Aug 2026 4 min read What I Learned Shipping Computer Vision Models to 10+ Edge Targets Notes on the gap between a model that trains well and a model that runs on someone's phone, car, or camera — quantization, runtime quirks, and why the bottleneck is rarely the network. Edge AIQuantizationDeploymentTensorRT Read the post