GeekWire’s in-depth startup coverage tells the stories of the Pacific Northwest entrepreneurial scene.
Guardrails AI, a San Francisco-based startup that helps companies manage risks when building AI applications, announced a $7.5 million seed round.
The company’s open source platform lets developers build and reuse validation techniques for their AI models. The idea is to secure reliability and prevent unintended outputs when working with large language models and other AI-related apps.
Founded last year, Guardrails is led by CEO and co-founder Shreya Rajpal, a longtime AI engineer who built self-driving tech at Drive.ai and autonomous systems at Apple.
Seattle tech vet Diego Oppenheimer is also a co-founder of Guardrails. He previously co-founded Seattle machine learning startup Algorithmia, which was acquired by DataRobot in July 2021.
“We’re on a mission to empower AI developers to build safer, more trustworthy AI applications,” Oppenheimer wrote on LinkedIn.
Oppenheimer is also a partner and CEO-in-residence at venture capital firm and incubator Factory, which invested in the seed round for Guardrails.
Zetta Venture Partners led the round. Other backers include Bloomberg Beta, Pear VC, and GitHub Fund, along with several angels.
Guardrails AI stands as the foremost open-source framework for defining and enforcing assurance in LLM (Large Language Models) applications.
It provides:
– A framework for crafting custom validations tailored to application needs.
– Orchestration of prompting, verification, and re-prompting processes.
– A library containing commonly used validators suitable for various use cases.
– Specification language facilitating the communication of requirements to LLM.