Ora Computing
Ora Computing is an AI deployment company specializing in optimizing and compressing Artificial Intelligence (AI) foundation models. They develop AI techniques that automate manual labor mostly on edge devices, such that small and performant models can be deployed on phones, laptops, IoT devices, and cars
xista team
Making advanced AI accessible across industries
Artificial General Intelligence is quickly transitioning from science fiction to reality, driven by astonishing advancements in Large Language Models (LLMs). Contemporary LLMs contain hundreds of billions of parameters and their scale is further expanding at a rapid pace. Model compression techniques have largely stabilized since the introduction of a few key techniques in the small model era, but carrying these over to the foundation model regime has proven costly, with teams of machine learning engineers spending months throwing different techniques at the wall to meet their device constraints.
Ora Computing’s innovative techniques drastically reduce parameter counts with minimal accuracy loss, outperforming traditional methods. They develop an automated deployment pipeline, enabling companies to productionize models in-house and monitor, debug, and adjust performance in real-time.
The Ora Team
Ora Computing was co-founded by a team of experts in machine learning with roots in physics. Their team possesses deep technical experience forged at top research institutions like Harvard, ETH Zurich and ISTA and top industry venues such as Apple, Citadel Securities, IBM, QuEra, and Pasqal.
Our Investment Rationale
With the increasing demand for efficient and private AI solutions, Ora Computing is poised to lead in deployment technology, making advanced AI accessible across industries.