Oumi, a Seattle-based startup, has raised $10 million in seed funding to launch an open-source AI research and development platform.
The company aims to open up the “black box” of foundational AI models, helping to foster innovation by making AI research more collaborative and accessible.
A “Linux Moment” for AI
The co-founders, Manos Koukoumidis (former senior engineering manager at Google Cloud) and Oussama Elachqar (former machine learning engineer at Apple and Twitter), believe AI needs its own “Linux moment”—similar to how open-source technology transformed computing.
The platform will support various AI tasks, from pretraining to inference and evaluation, enabling users to work with open models and commercial APIs.
Fostering Innovation with Open Collaboration
According to Koukoumidis, the restrictive nature of proprietary AI models from companies like OpenAI highlights the need for more open-source options.
Oumi plans to generate revenue by offering enterprise-tailored solutions, such as hosted training and inference workflows.
The company sees DeepSeek, a Chinese startup’s AI model, as a validation of Oumi’s strategy, as it utilizes open-source tools like PyTorch and Meta’s Llama.
Partnering with Top Research Universities
Oumi has partnered with 15 leading AI research universities across the U.S. and the U.K., including the University of Washington, to build the world’s largest AI research lab.
The aim is to lead AI development through open-source and transparent practices, focusing on creating trustworthy and explainable AI models.
Oumi is a public benefit corporation, indicating its commitment to social impact alongside profitability.
