A note

Where your data lives.

Aevum trains a shared base model on Dawn, the AIRR supercomputer at the University of Cambridge, using data Aevum itself has curated. That base model is the thing that ships. From there, a local configuration stays in the building and learns from how the buyer uses it.

What stays with you

Customer inputs, customer outputs, and the local learning the system does in response. These do not leave the building.

What we train on

Aevum's own curated training corpus. Not customer data. Not silently. Not opt-out.

What we will not pretend

"No training on customer data" is not the same as "no training". A model that never learns is a model that goes stale. The position here is that the learning happens locally, and stays in the building with the customer.