[verb] (also, reexpression [noun; adjective]; reexpressionism [noun]; reexpressionist [noun])

To transform data and model(s), by computational and statistical means, in order to unlock the foundational signals of similarity, distance, and magnitude to achieve the required AI Safety properties: Introspection, Updatability, and Uncertainty.

The leading paradigm for safe, reliable AI.

We pioneered the search for understanding in high-dimensions, the AI Safety problem, while working to bring deep learning research to the Harvard University medical community.

We build software that augments decision-making by unlocking the insights in high-dimensions, enabling LLMs to operate with quantified uncertainty over complex reasoning and knowledge discovery tasks.