About this blog

Cloudglue Engineering Blog is where we write about what we're working on at Cloudglue — the research we're digging into, and the technical learnings to share with the broader community.

We're building Cloudglue to make video a first-class citizen for AI systems. That means a lot of time spent on topics at the frontier of video understanding, and building infrastructure that scales to support the hundreds of developers and orgs that rely on us. And to scale to processing hundreds to thousands of hours of video per org.

We are capturing and sharing all of our learnings here.

What you'll find here

Research notes. Experiments with new models, comparisons of segmentation strategies, evaluations that surprised us, and the occasional dead end that taught us something worth sharing.

Product engineering. How we design APIs around messy video data, why we made a particular architectural bet, and what we've learned from developers building on top of Cloudglue.

War stories. The outages, the weird edge cases, the file formats that shouldn't exist but do. The stuff you only learn by shipping.

Who writes it

The technical team at Cloudglue. We are writing what we'd like to read ourselves — technical enough to be useful, honest about what didn't work, and short enough to finish over a coffee.

Stay in the loop

Subscribe to get new posts in your inbox.

Feedback?

We'd love to hear from you. Reach out to us at support@cloudglue.dev