float

We’re a team of mathematicians building advanced embeddings for financial market data. Our mission is to bring stability to global markets and uncover new insights at the frontier of generative AI.

What are embeddings?

Embeddings are a core concept in generative AI. They transform complex inputs—like an image, a paragraph of text, or a sequence of market events—into mathematical representations that capture their underlying meaning.

How can embeddings be used?

Embeddings unlock a wide range of applications from forecasting price movements and detecting illegal market manipulation to compressing data for more efficient storage and transmission. In true generative AI fashion, they can also be used to synthesize realistic market scenarios for simulation and stress-testing, especially rare “black swan” events.

Why the name?

In computer science, a float is a fundamental data type used to construct embeddings. The term also evokes movement and uncertainty: the path of a butterfly, or the price of a security buoyed by shifting demand.