Distributed Cloud for retail and manufacturing
summary
Demo shows an on‑prem inference stack built on Google Distributed Cloud using TPUs and a three-node managed environment that Google operates from firmware through Kubernetes. Models train in GCP and deploy to the edge for vision tasks like price‑tray item validation and manufacturing QC, claiming sub‑millisecond latencies for real‑time actions. The demo runs locally, scales to thousands of sites, and validates that cloud‑trained models can execute low‑latency inference at the edge.
AI-written summary · about
chapters (3)
- 0:00Intro
- 0:20Hardware-verse floor walkthrough
- 1:11Modern retail & manufacturing experiences on-premises
related talks
08mMaximizing ROI Through Distributed Computing and Multi-Cloud
23mBuilding Scalable, Observable MLOps Systems on Google Cloud
37mEnd-2-End Machine Learning on GCP
04mState-of-the-art Machine Learning research with Google tools
01mGoogle Cloud Platform at GoDataFest 2019
25mAn Introduction To Google Cloud Platform
26mBuilding a Cloud Native Edge Platform for Accelerating AI Inference
32mMachine Learning with Google Cloud Platform