Activeloop

Activeloop provides a database infrastructure platform, Deep Lake, designed for artificial intelligence applications that handle large-scale unstructured and multi-modal data. The company’s mission is to simplify AI data management and connect complex data like images, audio, video, and text to machine learning models.

The company, originally named Snark AI, was founded in 2018 by CEO Davit Buniatyan, Sergiy Popovych, and Jason Ge. Buniatyan’s background in neuroscience research at Princeton University, where he managed petabyte-scale imaging data, directly informed the company's direction. The founding team combines experience from institutions like Stanford, Google, and Tesla. Activeloop is an alumnus of the Y Combinator accelerator and is headquartered in Mountain View, California.

Activeloop has raised approximately $20 million in total funding over several rounds, including a $11 million Series A in March 2024, with investors such as Streamlined Ventures, Y Combinator, and Samsung Next.

The company’s flagship product, Deep Lake, functions as a data lake that fuses the flexibility of a data lake with the capabilities of a vector database. It is engineered to manage complex data by storing it in a columnar, tensor-based format, which enables rapid streaming to GPUs and deep learning frameworks like PyTorch and TensorFlow. This architecture boosts AI engineering productivity and can reduce costs. Key features include data versioning, advanced querying, vector search, and the ability to perform hybrid searches on embeddings and their metadata. More recently, Activeloop introduced Deep Lake PG, which unifies a serverless Postgres for transactional queries with Deep Lake's tensor storage for analytical workloads, positioning it as a unified database for AI agents.

Activeloop operates on an open-core business model, offering its core technology as an open-source library while providing enterprise-level subscriptions and licensing for its managed Deep Lake platform. The company targets enterprises, including Fortune 500 companies in sectors like biopharma, automotive, and MedTech, that are heavily invested in machine learning with large, unstructured datasets. Clients include Bayer, Matterport, Intel, and Google.

Keywords: Database for AI, Deep Lake, multi-modal data, unstructured data, vector database, data lake, AI agents, machine learning, data infrastructure, data versioning, GPU streaming, tensor storage, data pipelines, Davit Buniatyan, Y Combinator, AI development, LLM, retrieval-augmented generation, embedding database, computer vision, data management, MLOps, deep learning, enterprise AI, serverless database