Oumi

A platform for creating custom AI models in hours, not months. Automating evaluations, data synthesis, and training to create small language models (SLMs) that are more accurate and lower cost than out of the box models.

Oumi operates in insurtech based in Silicon Valley, California.

Profile

Oumi is a platform for custom AI models in days, not months – delivering up to 10× lower production costs and more accurate, reliable applications. As use cases hit production – especially agentic workflows – token usage and latency requirements spike, and real-user demands exceed PoC bars. Custom models (finetuned SLMs) are the answer, but many teams think they take too much time to develop. Oumi slashes that effort, enabling task-specific models that run at a fraction of the cost, keeping performance predictable, and avoiding lock-in to external APIs that can change silently. Oumi does it by using techniques to automate evaluations, data synthesis, and training. Founded by leaders who led Gemini and Cloud AI teams at Google, AI efforts at Apple, and Product teams at Cohere, Oumi pairs deep production experience with an open approach. Its open-source stack is used by engineers at Google, Microsoft, Adidas, and others, and its enterprise platform is available to select partners in early access.

As a business in insurtech, Oumi is listed here with the public details available about it.

The company is recorded as being based in Silicon Valley, California.

Its current recorded stage is Plug and Play accelerated startup.

Website: oumi.ai