Activeloop Raises $5M to Build Streaming Database for AI Models
Activeloop secures a $5 million seed round to launch a streaming database that helps developers efficiently feed unstructured media data into machine learning models. The system works like Netflix for AI, eliminating the need to download massive files.
Activeloop raises a $5 million seed investment led by 468 Capital and CM Ventures to build a specialized database for media-focused artificial intelligence applications. The startup, which emerged from Y Combinator's summer 2018 cohort, officially launches an alpha version of its commercial product today. This technology provides a dedicated storage layer that efficiently organizes and streams unstructured data like images, video, and audio directly into machine learning models.
The platform offers an open source API that converts complex media files into mathematical representations that machines understand, while also tracking different data versions and storing everything in repositories like Amazon S3. Instead of forcing data scientists to download massive files to a local machine, Activeloop streams the information to the AI model in the same way that Netflix streams video to a viewer. Founder and CEO Davit Buniatyan develops this streaming approach after struggling with enormous files during his neuroscience research at Princeton.
The open source project already boasts significant traction with 55 contributors, 700 community members, and 300,000 total downloads. With 15 current employees, Activeloop actively seeks to hire a half-dozen high-end engineers and intentionally prioritizes diverse talent from around the world. As the company grows its team and product, it aims to solve the massive data bottleneck that currently hinders computer vision, audio processing, and natural language processing applications.