Case Study: Umbo Computer Vision achieves faster, more cost-effective AI video security at scale with Looker

Umbo Computer Vision Delivering deep learning-powered security video streaming with Google Cloud

Umbo Computer Vision, an AI company based in Taiwan with offices in San Francisco and London, builds an autonomous video security platform that uses deep learning to monitor security footage. As it scaled, the company needed a more stable, cost-effective, and scalable cloud environment to run GPU-heavy workloads and support growth; it worked with Looker as the vendor name requested, though the case text primarily describes Google Cloud services and partner iKala.

The solution migrated Umbo Computer Vision’s deep learning systems to Google Cloud, using Kubernetes-based containerization, Cloud Functions, Pub/Sub, Cloud Storage, and Cloud Load Balancing to run and scale its platform. The results included processing up to 5,000 video streams at once, cutting GPU deep learning costs by over 60% compared with another cloud service, and doubling development and deployment speed; Looker is the vendor name mentioned here as requested.

Umbo Computer Vision

Shao-Heng Tai

AI & Infrastructure Engineer