Algorithmia offers a cloud-agnostic machine learning DevOps platform that lets data scientists train models with the framework of their choice, bring them to a preapproved enterprise platform, and deploy them to production. The company began five years ago as a marketplace for algorithms but has since pivoted to focus on helping enterprises productionize ML and AI. Its platform is used by more than 90,000 engineers and data scientists and counts the United Nations, several U.S. intelligence agencies and Fortune 500 companies among its customers. Algorithmia says CIOs are bogged down by the difficulty of managing and deploying ML models, and its tools target the complete machine learning lifecycle. With the new funding the company plans to invest in product development and build out its team, aiming to double headcount over the next year. CEO Diego Oppenheimer emphasizes reducing infrastructure burden so data scientists can focus on building ML and AI applications that deliver value. Algorithmia operates a marketplace and enterprise solution that gives developers access to 3,500 algorithms, functions and machine-learning models drawn from university researchers and individual contributors. Its CODEX platform is portable and runs on AWS, Azure and Google Cloud Platform, with support for private OpenStack deployments for regulated customers. The company has focused on security to safely run users’ arbitrary code, which has helped it win enterprise customers in finance, telecom and even some intelligence agencies. Algorithmia currently has about 45,000 developers on its platform and has begun growing its enterprise business by offering private, secured cloud deployments. The team is small today (about 13 people, mostly engineers) but plans to expand to roughly 23 staff by year-end and open a sales/solutions office in New York City. Algorithmia builds a platform and marketplace to give algorithm developers in academia and elsewhere a way to share their work. Uploaded algorithms are adapted to Algorithmia’s inputs and outputs and exposed via a REST API so developers can call them from any language. The company runs computation on its own instances that dynamically boot up across different cloud providers and locations to provide better performance. Users pay a base computation fee plus a fee to the algorithm’s creator if that creator chooses to charge, creating incentives to publish and improve algorithms. Algorithmia allows "black box" offerings but encourages contributors to publish source so the community can spot bugs and provide fixes. The service is currently in a private, invitation-only beta with a public beta planned for this fall, and the company says it will use recently raised capital to hire engineers, build its algorithm library, and add more cloud platforms.

Total raised

$38M

Funding rounds

3

Latest round

Series B

Latest activity

May 2019