turbopuffer - Fast, scalable vector and full-text search built on object storage
NablaTensor - Adjoint automatic differentiation engine on the JVM
Compute Cheap - The world's most affordable NVIDIA H100/H200 GPU compute capacity.
