AI models are a new type of tool. When we're given new tools, like the internet or electricity, especially so many so quickly, it takes us time to figure out how best to use them.

Imagine putting someone in front of a terminal to the internet in the 90s and expecting them to find it useful. We think AI is in a similar state, albeit as a set of tools more useful on their own than their predecessors.

Rather than viewing AI models as end-products, we view them more like a new building block for solving previously impossible problems.

At Sieve, we want more developers to think about AI models in this way so we're building a platform centered around this idea. Concretely, our platform is infrastructure for running AI models, a library of useful models that can be imported like Python packages, and a set of applications built on top of these models to solve various end use cases.

We think it'll drastically accelerate how developers solve problems with AI and the rate at which those solutions will be used by the rest of us. And video is just the start.
We're a team of engineers, scientists, and designers focused on simplifying today's AI ecosystem, starting with the video and audio domain. We come from top research labs and organizations, and almost every AI subdomain imaginable from AI hardware, to robotics, augmented reality, sports analytics, self driving, and more. Today, our technology is used by a wide variety of startups and enterprises who are leveraging AI to automatically understand or generate video and audio.