link your
corpus
generate
QA pairs
configure
rewards
automated by castform,
customizable by you
RL training
on your data
managed infrastructure,
full observability
model that
searches well